Source status: Public event pages, speaker biographies, conference archives, and linked media were checked. Recording and access status is marked separately for each surface.
Evidence Boundaries: A programme establishes that an organizer listed a speaker or topic. It does not establish attendance, a recording, current employment beyond the dated page, deployment, investment authority, or results.
Latest source refresh: The September 2 conference-watchlist note adds a canonical CEPR programme URL and the upcoming Neudata San Francisco Data Summit. Both remain dated programme surfaces, not evidence of implementation.
Additional September 2 route: QuantInsti Annual Algo Trading Conference 2026 advertises a September 24 online programme built around a live agentic strategy workflow, intentional failure testing, HFT infrastructure, and changing quant roles. Speakers had not yet been announced when checked; the page is therefore a capture queue, not a named-firm disclosure.
The conference universe is larger than the currently verified application routes. The public record is most useful when each event is described by its speaker evidence, audience, access route, and unresolved capture question:
- A current AI owner or practitioner from a relevant fund has already spoken.
- The organizer publishes a speaker, nomination, CFP, or practitioner route.
- The audience includes buy-side technology, research, risk, or data leaders.
- The event has a clear location, access, or relationship context.
- The topic can be presented without disclosing proprietary systems or returns.
These are descriptive checks, not a scoring system or event ranking.
Conference and application surfaces
| Surface | Public evidence | Application or capture route | Open question |
|---|---|---|---|
| AI Loves Data New York 2026 | Public event and speaker/attendee surfaces are linked in the source list. | Verify the current date, speaker route, and finance-specific sessions. | The public page has had date ambiguity; confirm before treating it as current. |
| GAIIM 2026 | The official page advertises a September 29–30, 2026 hybrid event at Columbia University’s Faculty House and lists Pete Petersen (Causeway Capital), Michael Soss (Millburn), and George Ho (NIV Asset Management), alongside finance-data and AI vendors. The preliminary programme includes agentic buy-side workflows, Claude Code/Cursor, MCP-connected research stacks, governance, and an Anthropic financial-services presentation. | Recheck after the event for public recordings, slides, speaker posts, sponsor material, and transcripts; capture each session separately. | The page establishes advertised programme and listed roles only. Attendance, session content, internal systems, permissions, and results remain unresolved. See the capture note. |
| CFA Society Hong Kong AI x FinTech Symposium 2026 | The official page advertises a September 17, 2026 Hong Kong symposium and lists Kevin Kwan (Bloomberg Enterprise Data), Douglas Chan (eBroker), Bryan Chik (BlackRock), Alice Wong (J.P. Morgan Asset Management), and Brooksley Kang (Aletheia Capital), with sessions on AI investment workflows, agents, infrastructure, and digital assets. | Recheck after the event for recordings, slides, speaker posts, and session-level material; separately cross-check each role against first-party employer pages. | The page establishes advertised programme and listed roles only. Attendance, session content, internal systems, permissions, and results remain unresolved. See the capture note. |
| MFA Ops: Data, Tech & AI | MFA provides an alternative-manager operator route. | Check peer-roundtable, moderator, and speaker pathways. | Attendance and participant details may be gated. |
| Georgia Tech AI and the Future of Finance | The 2026 programme connected finance-AI practitioners with the Georgia Tech financial-services research ecosystem. | Track the next edition, guest lecture, panel, or QCF route. | The next dated programme and recording status require verification. |
| AI Engineer NYC 2026 | The public event material includes an AI-in-finance speaker route and technical agent-engineering content. | Verify CFP status and finance-track recordings. | Current deadline and final programme require confirmation. |
| CFA Society Netherlands — AI, Machine Learning and Agentic Investing | The June 4, 2026 programme names Robeco, Northern Trust Asset Management, Argan.ai, and TransTrend speakers across human-AI investing, ML in quant equities, agentic macro investing, and trend research. Argan’s public post corroborates the speaker map, and a post-event speaker summary frames general-purpose models as potentially amplifying familiar quant problems when inserted into the research process. | Search the members-only archive, speaker pages, and public reposts for recordings or slides. | Public programme and speaker commentary are available; no public recording, slide deck, model inventory, data-permission map, or investment result was recovered. |
| MathFinance Conference 2026 | The September 10–11 programme names XTX Markets/Oxford-linked researcher Hans Bühler, QuantZ CIO Milind Sharma, AI for Alpha, and a session on AI-augmented model validation. | Track the promised video/audio recordings and speaker materials after the event. | Recording publication and access conditions are not yet public. |
| Imperial–Goldman Sachs–CEPR Hedge Fund Conference | The completed 1 July 2026 programme names Goldman Sachs, Imperial/CEPR, academic, QRT, Robeco, Allspring, and other participants. Sessions include “Machine Learning Meets Markowitz,” narrative momentum, thematic investing, and statistical arbitrage/machine learning. | Search the recovered programme, post-event organizer updates, speaker papers, and any CEPR/Imperial media for recordings or slides. | The reviewed public surfaces establish dated programme and contributor evidence, but not recordings, slides, model inventories, live deployment, permissions, or results. |
| QuantMinds Quant Invest Summit | The November 16–19, 2026 London event lists quant-fund and asset-manager speakers including Fidelity International, Assenagon, and Edo Theory. | Track the agenda, speaker biographies, and any post-event media. | AI-specific sessions and recording availability remain unresolved. |
Newly recovered conference surfaces — August 18, 2026
| Surface | Public evidence | Capture boundary |
|---|---|---|
| Future Alpha 2026 | The March 31–April 1 New York agenda lists Bryan Kelly (AQR), George Patterson (PGIM), Jacob Bowers (BlackRock), Frank Ieraci (CPP Investments), Andrew Gelfand and Charlie Flanagan (Balyasny), John Feminella (Two Sigma), Dhrupad Bhardwaj (Balyasny), Barry Fitzgerald (Man Group), Mike Chen (Robeco), Alan Russell (Brevan Howard), Irina Bogacheva (Millburn), and Haoxue Wang (Millennium), among others, across ML, AI, alternative data, execution infrastructure, alpha capture, and front-office engineering sessions. | Third-party post-event agenda; recordings, attendance, and substantive session content remain unresolved. Promotional or unsupported performance claims on the page are excluded. |
| CQF Institute AI and Machine Learning in Quant Finance Conference | The September 16, 2026 online event lists Petter Kolm, Yuyu Fan of AllianceBernstein, Tony Guida, Hariom Tatsat, Jörg Kienitz, and Lorenc Kapllani. The page exposes university lineage and research topics, including LLM-derived financial-news networks, graph-neural-network covariance construction, knowledge graphs, and systematic investing. | Upcoming programme and speaker-bio evidence; future recordings and slides are not yet captured, and biographies do not establish live production systems or results. |
Open-source and allocator/data conference routes — August 28, 2026
The EQDerivatives Global EQD 2026 programme adds a systematic-investing and derivatives conference route. The Las Vegas event took place May 20–21, 2026; its programme lists Milind Sharma of QuantZ Machine Intelligence Technologies for an academic keynote on The Quantamental Revolution — Factor Investing in the Age of Artificial Intelligence. It also displays public personnel routes for Akshay Padmanabha (Balyasny), Karen Luong (Man AHL), Arnab Sen (Paloma Partners), Vishnu Kurella (a Millennium platform company), and Mayank Gupta (Fidelity), with sessions on machine-learning volatility prediction, systematic strategies, portfolio construction, and QIS infrastructure. This is programme and speaker-network evidence, not evidence of attendance, shared systems, adoption, or performance.
The same organizer lists Asia EQD 2026 for October 27–28, Investor EQD — Abu Dhabi 2026 for November 18–19, and Japan EQD 2026 for November 30. These future pages are conference-watchlist surfaces for Asia-Pacific, Middle East, and Japan systematic-investing coverage; named speakers, recordings, and session-level AI evidence still need separate verification.
The Asia EQD agenda adds session-level detail to that watchlist. Its October 27–28 Hong Kong programme includes “Enhancing Portfolio Optimization — AI and Investment Decisions,” with April Fu of Ping An of China Asset Management Hong Kong and Daniel Xystus of AOP Capital listed as panelists. It also names Kaushik Ramakrishnan of Jain Global, Oliver Chan of Capula Investment Management, David Dredge of Convex Strategies, Darren Kang of Life Asset Management, Mengyun Zhang of China Re Asset Management, and other regional investment personnel. The event is upcoming; no attendance, recording, transcript, model, or firm implementation is inferred. See the capture note.
Two additional official event pages expand the title-blind conference queue. The Open Source Quantitative Finance conference (osQF, formerly R/Finance) is scheduled for October 23–24, 2026, at the University of Illinois Chicago, with a PyData Chicago kickoff on October 22. Its stated topics include AI, risk, econometrics, high-performance computing, market microstructure, portfolio management, and time-series analysis in the context of open-source software for model development, portfolio construction, risk management, and trading. The page displays a committee with links to Petra Bakosova, Kyle Balkissoon, Oleg Bondarenko, Brian Peterson, Dale Rosenthal, Jeffrey Ryan, Justin Shea, and Joshua Ulrich. These are organizer and affiliation routes, not evidence of a covered-firm presentation, deployment, or result.
The Battle of the Quants London 2026 page dates its event to October 22, 2026, at the Royal Automobile Club. The organizer describes an audience of allocators, portfolio managers, data buyers, and data providers; lists AI, ML, LLMs, agentic AI, and alternative data as discussion areas; and advertises invitation-only Allocator Breakfast and Data Breakfast sessions. The accessible page exposes a preliminary agenda image but no named speakers in its text. Its official videos and past-events navigation create a post-event recording-recovery route. No attendance, speaker participation, dataset use, model deployment, or performance claim is inferred. The full capture note is here.
The organizer’s official videos page adds historical, role-labeled attendee comments: it displays “Vice President, Two Sigma (London),” “Head of Data Science, Berenberg (London),” a “Deputy CIO, California-based Hedge Fund,” a “CEO & CIO, Canadian-based Hedge Fund,” a “COO, Hong Kong-based Quant Fund,” and a “Director, Alternative Data Provider,” among other anonymous labels. The page does not name these individuals, date each comment, attribute it to a particular session, or provide a transcript. They are participant-discovery seeds, not evidence of 2026 attendance, a named person’s identity, a firm’s system, data purchase, or performance.
The Oraclum Capital LinkedIn announcement adds a separate New York City route: it announced an April 28, 2026 panel titled “From Models to Markets: AI’s Role in Quant Trading,” naming Vuk Vukovic of Oraclum Capital, Evan Schnidman of Fidelity Labs as moderator, Yves Lemperiere of CFM, and Ziang Fang of Man Group. A public organizer post names additional announced participants from Crabel Capital Management, Lighthouse Investment Partners, Engineers Gate, Schonfeld, Columbia University, ClearAlpha Technologies, and AbleMarkets. The linked New York event page now returns 404, and no authorized recording or transcript was found. These posts establish an announced event lineup and topic scope, not attendance, implementation, or performance. The capture note records the route and its boundary.
The organizer’s homepage also contains two “Previous Event Videos” YouTube embeds, labeled Hong Kong (3kpCBlVcsrc) and London (AfeRpt9RvSw). Both direct locators returned unavailable through the permitted recovery path on August 28, 2026, so no transcript or session claim is promoted. The embeds remain explicit retry targets and are recorded as failed captures rather than omitted.
| Imperial–Goldman Sachs–CEPR programme PDF and post-event organizer update | The recovered agenda provides the session-level speaker and discussant map. The post-event update confirms that the event took place on 1 July 2026 and identifies the presentation/Q&A contributors. The public topic surface spans machine learning in portfolio construction, narrative asset pricing, thematic investing, statistical arbitrage, and options. | Resolve speaker papers, institutional reposts, and any recording or slide archive. Maintain separate dated associations for Goldman Sachs, QRT, Robeco, Allspring, and academic participants. | No recording, slide deck, model inventory, permission map, live fund deployment, or performance evidence was found in the reviewed public surfaces. |
| Neudata San Francisco Data Summit 2026 | The October 29, 2026 agenda lists sessions on AI for fund managers, alternative-data sourcing, compute pricing, entity resolution, textual/audio/social inputs, MCP access patterns, and AI workflows for investment firms. Displayed participants include Neudata, Integrated Quantitative Investments, Blue Owl, Perplexity, Silicon Data, Senzing, and N47 personnel. | Recheck after the event for session recordings, slides, speaker posts, buyer identities, and transcript routes. | The page is an upcoming programme. It does not establish attendance, buyer/vendor relationships, data licensing, implementation, or performance. |
| CFA Society Netherlands event page and Argan public event surface | The dated programme is now corroborated by public company/speaker posts. The accessible post-event framing emphasizes that a general-purpose pre-trained model can magnify known quant problems and that research-process design precedes model evaluation. | Search the members-only archive and named speakers’ public pages for slides, recording, or a longer written distillation. | This remains programme and speaker commentary; it is not a transcript, model audit, live deployment disclosure, or performance study. |
Statistical Horizons / AI Horizons: an applied methods and research-agent surface — August 20, 2026
The supplied Statistical Horizons home page, events page, page two, and page three add a different kind of discovery surface. They are public training catalogues, not hedge-fund disclosures, but their title-blind archive exposes methods and workflow vocabulary that a search for “hedge fund AI” would not reliably find. Page two includes Machine Learning for Estimating Causal Effects, AI-enhanced audit and correspondence experiments, Claude-powered academic research, and longitudinal models with LLMs. Page three includes Getting Work Done with AI Agents, AI Tools for Data Analysis, and AI-Based Emotion Analysis.
The catalog also has a durable machine-readable route: the public WordPress REST root, RSS feed, events feed, and Event Espresso event endpoint. The August 20 capture returned 500 event records. These are event inventory, title, slug, date, and modification-time routes; no attendee, registration, contact, or answer endpoint was requested.
| Public lane | What the pages expose | How to interpret it |
|---|---|---|
| Causal ML and algorithm evaluation | Ashley Naimi’s causal-effects course names double-robust ML, TMLE, and AIPW. Kosuke Imai’s public instructor profile connects causal inference, program evaluation, heterogeneous treatment effects, individualized treatment rules, selective labels, and human–AI collaboration. | A useful validation vocabulary for research features and decisions. It is not evidence of a named fund’s implementation or result. |
| Applied ML and interpretability | Bruce Desmarais’s ML course covers cross-validation, model evaluation, variable selection, random forests, and XGBoost. Adam D. Rennhoff’s interpretable-ML course names PDP, ALE, ICE, LIME, anchors, SHAP, and uncertainty. Cynthia Rudin’s course adds sparse models, GAMs, Rashomon sets, and interpretable neural networks. | A public methods map for model auditing and error analysis, not a private model inventory. |
| Research agents and reproducibility | AI Horizons describes bounded delegation, review checkpoints, file-aware agents, logs, project memory, web search, multiple agents, API-error handling, and the loop “do work → log → learn → automate.” The AI research-workflows certification combines these with Claude, Codex, and R-based research workflows. | A workflow blueprint for inspectable automation. Training claims do not establish adoption by a manager or asset owner. |
| Document-to-data | Turn Documents into Data with AI teaches codebooks, structured extraction, evidence-linked cells, disagreement checks, and reliability. The Data Mint page describes multiple extraction passes and audit trails. | Relevant to filings, transcripts, clinical-trial documents, and other corpora. Data Mint’s scale, pricing, privacy, and provenance statements are vendor claims. |
| Text, embeddings, and RAG | Text Classification with LLMs in R names embeddings, zero-shot classification, Hugging Face Transformers, RAG, encoder models, network models, and time-series text. | Expands the search beyond generic “sentiment” to representation, retrieval, and structured text pipelines. No investable signal is established. |
| Multimodal analysis | The emotion-analysis course names BERT text models, CLIP image/video models, small-LLM rationales, multilinguality, class imbalance, frame sampling, and the MAFW dataset subset. | A lead for voice, image, and video research. It is not evidence of earnings-call deployment, deception detection, alpha, or returns. |
The Statistical Horizons about page describes AI Horizons as a 2026 division focused on AI/LLMs in research and data analysis. Public instructor pages connect the catalog to academic methods at Harvard, Duke, Penn State, UVA, and other institutions, including Bruce Desmarais, Kosuke Imai, Cynthia Rudin, and Hudson Golino. Those are lineage and methods routes, not personnel links to hedge funds.
The AI Skills article
is also relevant to this project’s ingestion design: it describes
SKILL.md packages containing instructions, scripts, templates, and references
for file-aware agent workflows. This corroborates a public, inspectable skills
pattern, but does not independently validate its effectiveness.
The full capture, archive route, instructor map, and evidence boundaries are recorded in the Statistical Horizons source note. The practical implication is a better recovery rule: search methods catalogues and instructor ecosystems for causal validation, interpretable models, document extraction, embeddings, multimodal evaluation, and bounded agents; then resolve the cited paper, package, dataset, license, revision, and evaluation before connecting anything to a fund.
CQF and QES agentic-finance surfaces — August 19, 2026
The title-blind catalogue pass found a separate Quantmate/CQF route and a QES AI-in-Finance programme that were not in the media ledger. The CQF industry talk describes a May 14, 2026 online session on in-context reinforcement learning and test-time compute and identifies Nicole Königstein with Quantmate. The Quantmate company page describes agent populations that code, explain, and evolve strategies under live evaluation; this is a company-profile claim, not proof of a regulated fund, customer deployment, or live trading authority.
The QES AI in Finance programme adds dated speaker/topic evidence for LLM knowledge graphs and systematic thematic investing, earnings-call Q&A, LLM-based core-earnings measurement, agentic LLMs, AI-agent collaboration in financial research, and multimodal financial LLMs. It names QES/Wolfe Research, Boston College, Harvard, Quantmate, Maryland, Rutgers, and Columbia participants. No recording, slides, model inventory, portfolio permissions, or investment result was captured. The full capture note is here.
An exact-title search recovered a Cornell CFEM and UBS seminar page and a Bayes Business School workshop listing for the same Quantmate talk. Cornell’s abstract explicitly names multi-agent systems, LangGraph, and Language Agent Tree Search. These are university and workshop programme surfaces; they add architecture vocabulary but no recording, customer, permission, or performance evidence.
The Wolfe Research events calendar also confirms three current QES-related surfaces: the 8th Annual AI in Finance Conference in New York on March 24, 2026, the 3rd Annual Canadian Quantitative and Macro Investment Conference in Toronto on May 5, 2026, and the 9th Annual QES European Quant and Macro Investment Conference in London on May 14, 2026. These are official calendar records and recovery routes, not evidence of attendance, recording access, or session substance.
Wolfe’s public “Seven Sins of Investing with GenAI” webcast page has June 2, 2026 page metadata, but exposes only a title and registration form in the public response. The Wolfe LinkedIn announcement confirms the March 2026 AI-in-Finance event. Neither surface supplies a public transcript, speaker list, recording, model inventory, or investment result. The capture note records these as durable discovery leads rather than substantive system disclosures.
Wolfe Event Smart archive: a longitudinal QES research trail
The six linked Wolfe Event Smart pages add an archive layer that is not visible from a current-event search alone. The Event List currently returns no active event entries, while the Venues index preserves a recurring New York, Toronto, and London footprint. The venues show Nomura Securities and Wolfe Research offices in New York, Toronto hotels, and the May Fair and Oxford & Cambridge Club in London. This is organizer-history evidence, not attendance or sponsor evidence.
The linked pages and their agenda PDFs expose the following evidence lanes:
| Wolfe/QES surface | Publicly stated material | Boundary |
|---|---|---|
| 7th Annual AI in Finance, March 27, 2025 and agenda PDF, V5 | Yin Luo and team host a Nomura event with speakers from QES/Wolfe, Boston College, Penn State, Harvard, Quantmate, Maryland, Rutgers, and Columbia. Topics include earnings-call Q&A, LLM-assisted research, dynamic knowledge graphs, core-earnings measurement, agentic LLMs, AI-agent collaboration, GenAI in conference calls, continuous auditing, and multimodal financial LLMs. A separately reachable older agenda object lists a different speaker for one forecasting slot. | Dated programme and speaker/topic evidence. No public recording, code, model inventory, portfolio permission, or result. Preserve agenda-version differences and resolve attendance only through speaker CVs or authorized media. |
| 7th Global Quantitative and Macro, October 18–19, 2023 and agenda PDF | The second day includes Harry Mamaysky on news, Alex Kim on extracting corporate risks from transcripts with generative AI, Nick Guest on indirect disclosure, and adjacent sessions on tail risk, estimation risk, factor-zoo predictability, and sentiment risk premia. | Research questions and speaker affiliations, not evidence that Wolfe, a client, or a named fund traded the signals. |
| 8th Global Quantitative and Macro, October 23–24, 2024 and agenda PDF | The intervening programme names “Triple Sentiments” combining analyst ratings, textual NLP sentiment, and speech emotions; AlphaManager’s data-driven robust-control approach; LLM macro forecasting; ML for market microstructure; uncertainty-averse investors; and drug-development disclosures as alternative data. | Dated programme and speaker/topic evidence. It does not disclose source data, code, validation, permissions, or a named fund’s implementation. |
| QES job-postings luncheon, May 29, 2024 | The page says the presentation used RavenPack job-postings data covering more than 200 million postings from more than 60,000 companies in 195 countries since 2007, with more than 10,000 skillset metrics. It describes hiring, job-growth, geographic-expansion, staffing-demand, and technical-skill factors. | Wolfe/QES description of a research presentation. No license terms, point-in-time feature construction, backtest, or performance result is public on the page. |
| 8th QES European Quant and Macro, May 13, 2025 and agenda PDF | The London programme includes Sig-Wasserstein GANs for conditional time-series generation, graph-clustering statistical arbitrage, agentic and multimodal thematic investing, random-forest macro forecasting, LLM sentiment analysis with proximal-policy optimization, and LLMs in finance. | Dated programme evidence. The agenda does not establish implementation at any named fund. |
The AI-in-Finance landing page adds metadata that the agenda does not: Wolfe explicitly framed the March 2025 event around foundation models including DeepSeek R1, GPT-4, and Claude, and advertised a panel involving Chief AI Officers or senior practitioners from AllianceBernstein, Goldman Sachs, T. Rowe Price, and Cubist Systematic. The public page does not identify the panelists, disclose their remarks, or show which models any institution used. This is therefore useful as a title-blind organization and model-vocabulary lead, not as evidence of a named firm’s production architecture.
The archive also exposes a hidden 2019 QES NLP and Machine Learning event page and its agenda PDF. That programme names RavenPack, Accern, and AlphaSense vendor presentations; Wolfe/QES sessions on news, social media, corporate events, and filing text mining; and a media panel that included Bloomberg News’s head of AI and information extraction. It gives the research catalogue a dated starting point for the public progression from textual and news analytics to alternative data, LLMs, agents, graphs, and multimodal finance. The progression describes public topics and relationships. It does not establish comparative capability, production deployment, or investment results.
A separate 2019 Artificial Intelligence and Data Science in Trading delegate brochure adds a cross-firm speaker graph outside the Wolfe archive. It names Sebastien Guglietta as Brevan Howard’s then Co-Head of Computational Intelligence Systematic Strategies, Anthony Ledford as Man AHL’s Chief Scientist, Andrew Janian as Two Sigma’s Head of Data Engineering, Lisa Schirf as former COO of Citadel’s Data Strategies Group and AI Research, and Afsheen Afshar as former Chief AI Officer at Cerberus. The agenda themes include AI ethics, trust and interpretability, alternative-data alpha capture, GPU-accelerated AI/ML, and the future of quantitative research. These are dated conference affiliations and organizer descriptions; they are discovery routes, not evidence of current employment or deployment. The full capture note is here.
Historical speaker graph resolved against current sources
The 2019 brochure becomes more useful when its names are followed into first-party pages and dated research rather than treated as a current personnel list. The speaker-resolution note records the evidence classes and remaining gaps.
| Historical route | Public follow-up | What the sources support |
|---|---|---|
| Afsheen Afshar — Cerberus | Cerberus’ 2017 release and Columbia workshop | Cerberus described a proprietary AI/ML operations platform intended to support portfolio companies and trading desks; Columbia records enterprise-AI work on legacy systems, end-user design, and adoption. This is historical Cerberus evidence, not a current employment record. |
| Anthony Ledford — Man AHL | CFA/Man case study, Oxford-Man ML guide, and Man overfitting discussion | Public materials document the 2009 ML-specialist hire, an SVM prototype that rediscovered known relationships, later diversifying signal research, and a review/test-trading/client-eligibility process. These are dated Man Group materials, not a current model inventory. |
| Andrew Janian — Two Sigma | Current Citadel leadership page | Citadel currently identifies Janian as interim CTO and records 13 years at Two Sigma, including Head of Data Engineering and CIO. The page supports a current technology-lineage route, not ownership of a named AI model. |
| Lisa Schirf — Citadel | CMU panel record, Tradeweb interview, and Ai-Price release | CMU records the former Citadel data/AI role; Tradeweb later describes Schirf’s data-strategy leadership and a machine-learning pricing service covering approximately 880,000 municipal securities and custom portfolios. This is a vendor/product route, not current Citadel evidence. |
| Sebastien Guglietta — Brevan Howard | Laser Digital biography, Nomura/Laser announcement, and dated Brevan personnel report | Laser Digital currently identifies Guglietta as Head of Asset Management and records prior Nomura and Brevan roles. The Brevan AI-group evidence remains historical; a personal LinkedIn LLM experiment is not an employer disclosure. |
The cross-firm graph therefore yields current-personnel routes, historical research-process evidence, and vendor/product follow-ups. It does not support a ranking, a current model census, or a claim that any one firm is further along.
The API also preserves two easy-to-miss 2023 surfaces that the visible event list does not expose clearly. The AI Revolutionizing Investment Management and Transforming Everyday Life page records a one-day Wolfe conference hosted by Yin Luo and the QES team on May 22, 2023. Its public page gives the date, Wolfe office, and paper-and-panel format, but no agenda, speaker list, recording, or model detail. It is a personnel and recovery lead, not a substantive AI-system disclosure.
The September 19, 2023 QES luncheon on disrupting thematic investing is more revealing at the methodology-description level. Wolfe says the seminar covered AI, Wolfe Alpha List, cleantech, and disruptive-innovation thematic baskets; both long-only and long/short implementations; and a portfolio-construction technique intended to remove systematic style risk while preserving the original long-only book. The page also explicitly says the baskets account for risk, liquidity, cost of borrow, market impact, and transaction costs. This is a first-party description of a research/product surface, not a disclosed client portfolio, model specification, backtest, or performance result.
The agenda is also a useful title-blind personnel and vendor seed list. It identifies Armando Gonzalez (RavenPack), Kumesh Aroomoogan (Accern), Chris Ackerson (AlphaSense), Vojislav Maksimovic (Maryland), Alejandro Lopez-Lira (Wharton), and Monique White (Bloomberg News’s AI and information-extraction function), alongside Financial Times, Barron’s, and Risk.net journalists. Those names open distinct follow-up routes—vendor research symposia, product-history pages, academic papers, newsroom-AI profiles, and later conference appearances. They are historical programme affiliations, not evidence of a Wolfe purchase, hedge-fund deployment, or current employment.
The 2024 agenda expands the cross-cohort map beyond text and LLMs. Speech emotions appear beside analyst ratings and textual sentiment; macro forecasting appears beside LLMs and uncertainty-aware prediction; and the drug-development session creates a specific bridge to clinical-trial and regulatory-event research. Those associations justify targeted paper, author, and recording searches, but the conference programme alone cannot establish that any Wolfe client ran these signals in a live portfolio.
The agendas are paper indexes, not just speaker lists
The supplied pages become materially more useful when each distinctive agenda title is followed into the public research record. That pass recovered several specific model and data routes:
| Agenda title | Public follow-up | Evidence boundary |
|---|---|---|
| From Transcripts to Insights: Uncovering Corporate Risks Using Generative AI | Alex Kim, Maximilian Muhn, and Valeri Nikolaev use GPT-3.5 on earnings-call transcripts to construct political-, climate-, and AI-risk measures. Their paper reports information content for abnormal volatility and corporate investment/innovation outcomes, with risk assessments carrying more information than short summaries. | Academic transcript-risk research linked to the 2023 agenda. It is not evidence of Wolfe or hedge-fund deployment, permissions, or a post-cost trading result. |
| Enhancing Investment Analysis: Optimizing AI-Agent Collaboration in Financial Research | The authors describe configurable agent-group sizes and collaboration structures for fundamentals, market sentiment, and risk analysis on 2023 SEC 10-Ks from 30 Dow companies. | A concrete multi-agent research architecture linked to the 2025 agenda. The experiment does not establish live data access, monitoring, investment authority, or capacity. |
| Open-FinLLMs | The project describes FinLLaMA pretraining on a 52-billion-token financial corpus, instruction tuning on 573,000 financial instructions, and FinLLaVA tuning on 1.43 million image-text pairs across text, tables, time series, and charts. | A public multimodal-finance model route connected to the agenda’s Open-FinLLMs slot. Reported simulations and benchmark results require independent replication; they are not evidence of hedge-fund use. |
| LLM-based sentiment analysis for portfolio allocation with proximal policy optimization | This paper describes a sentiment-augmented PPO pipeline using daily Refinitiv news sentiment generated with LLaMA 3.3, evaluated on a three-stock portfolio. | A specified sentiment-to-policy experiment linked to the 2025 European agenda. The small study does not establish robustness, capacity, transaction-cost viability, or production use. |
| Scaling Core Earnings Measurement with Large Language Models | The HBS/Columbia summary connects the agenda title to extracting persistent or “core” earnings from 10-K disclosures. | A filing-to-fundamentals research route, not a disclosed Wolfe product implementation or client system. |
The practical implication is a better recovery rule: exact agenda title → author page → paper repository → code/data page → university seminar or authorized recording. The agenda itself establishes a dated public research association. The paper is where the data scope, model revision, evaluation split, and limitations become inspectable. The full paper-level capture note is here.
Wolfe EventSmart API archive: additional research routes — August 19, 2026
The supplied pages are only a slice of the archive. The public Event Espresso events endpoint returned 34 Wolfe event records dated 2018–2025. The visible Event List says its calendar feature is disabled, making the API, individual event pages, and linked agenda PDFs the more durable discovery route. The archive includes repeated NLP/ML, European, global, and virtual conferences as well as Asia, Japan, Canada, ESG, options, thematic investing, earnings risk, innovation, AI, and corporate-web research surfaces. This is archive evidence, not evidence of attendance or implementation.
The linked datetime objects add a separate, aggregate metadata layer. For
example, the API exposes DTT_sold values of 145 for the 2019 QES NLP event,
399 for the 2020 virtual NLP event, 425 for the 2021 virtual NLP event, 305 for
the 2023 global quant/macro event, 284 for the 2024 global event, and 310 for
the 2025 AI event. These are Event Espresso ticket/registration counters, not
attendance figures; the public API does not reconcile approvals, cancellations,
duplicates, or check-ins. They are useful for tracking event format and public
registration metadata over time, not for ranking conferences or inferring
research impact. The archive also exposes EVT_created, EVT_modified, and
exact datetime fields, which make it possible to detect later page or agenda
changes. The capture note documents the selected event IDs and boundaries:
Wolfe EventSmart API archive.
The supplied links also reveal an ingestion and privacy boundary. The disabled calendar is a WordPress page with a public REST representation, and the site REST root advertises Event Espresso namespaces for event metadata alongside attendee, registration, and answer resources. The capture uses event and venue metadata only; it does not query personal registrant data. The venue page’s public RSS route is useful for tracking venue additions or edits, while the general site feed currently contains no event items. Future sweeps should preserve the event ID, page slug, agenda URL, page/API modification time, venue record, and retrieval status as separate provenance fields. A discoverable endpoint is not, by itself, evidence that personal data should be collected.
There is also a recoverability trap. A read-only check of the 24 agenda-PDF URLs embedded in the API event descriptions returned HTTP 404 for the original EventSmart upload paths on August 20, 2026. Several of the agendas are still available through separately indexed CloudFront objects, but the page’s link itself is not a reliable asset-presence test. The capture ledger should record the event record, original asset URL, replacement asset URL, HTTP status, and retrieval date as separate fields. The detailed asset-recovery note is here.
Several records add specific signals to the public research map:
| Recovered route | Publicly stated research surface | Boundary |
|---|---|---|
| 2021 QES Innovation-Themed Seminar and agenda | Fintech, AI, systematic investing, data science, innovation alpha from ARK discretionary ETFs, and an innovation-in-data-science talk by Wolfe QES. | Dated programme and speakers; no code, recording, deployment, or result. |
| 2023 Harnessing Options and agenda | Options data framed for sentiment, price discovery, timing, volatility estimation, stock selection, earnings ideas, and risk-model factors; includes Wolfe QES, Nomura, academic, and NewMark Risk speakers. | Public research and vendor map; no licensed-data or portfolio-use proof. |
| QES earnings-risk luncheon | Wolfe describes an ERIE model combining earnings-growth expectations, earnings uncertainty, and a systematic earnings-prediction model, with a risk-model integration route. | Publisher description; no feature history, validation protocol, or performance table. |
| 2024 Investing in Japan and agenda | Japan alternative-data ecosystem, stock selection labeled alpha/machine learning, exposure measurement, BOJ policy, and QES risk models. | Regional conference evidence; no named external-fund deployment claim. |
| 2025 Canadian Quantitative and Macro Conference and agenda | “Fed Sentiment with LLMs,” machine-learning portfolio performance, generalized factor neural networks, the CRACKS corporate crash-risk model, deep-learning factor timing, and portfolio construction. | Dated agenda topics; no evidence that a named fund adopted a method or of resulting returns. |
| Pixels to Profits | A corporate-websites-to-stock-performance research route is preserved in the archive. | The current page exposes no substantive description or recording; treat as a recovery lead. |
The full archive note is here.
Wolfe QES product surfaces connected to the conference archive — August 19, 2026
The EventSmart links become more useful when paired with Wolfe’s own current product pages. The QES Data Feeds page names an NLP model family and describes alternative data, factor construction, machine learning/NLP, stock selection, ESG, and global macro research. Its public model descriptions include:
| Public QES surface | What the page says | Evidence boundary |
|---|---|---|
| GINA | Textual alpha sourced from regulatory filings, company call transcripts, 13Fs, Forms 3/4, and global filings | A public product description; no code, vendor contract, feature history, or independently audited result |
| BRAIN | AI-related patenting activity combined with NLP analysis of company filings and call transcripts | A strategy description; no disclosure of model architecture, portfolio permission, or implementation at a named outside fund |
| StarPerformer and NEMO 2.0 | Deep-learning mean-reversion language for StarPerformer; executive/board, supply-chain, geographic, shipping, momentum, and NLP-network inputs for NEMO 2.0 | Firm-reported model descriptions; performance language on the page is not an independent audit |
| PDI/CPDI | A 120-factor family using accounting variables to study anchoring and delayed reaction to new information | Public factor-description evidence; no point-in-time feature files or backtest protocol is exposed |
The investable-strategies page also states that QES has published more than 100 papers, maintains 30+ live models, tracks investable strategies daily, and reports them monthly. Those are Wolfe’s own product-page statements. They should be read as a map of public research and commercialization surfaces, not as proof of hedge-fund adoption, performance, or comparative capability.
The current official event calendar adds three 2026 discovery anchors: the 8th Annual AI in Finance Conference in New York on March 24, the 3rd Annual Canadian Quantitative and Macro Investment Conference in Toronto on May 5, and the 9th Annual QES European Quantitative and Macro Investment Conference in London on May 14. The corresponding conference pages currently redirect to Wolfe’s home page, so the calendar is a lead for agenda, speaker, and recording recovery rather than a recording itself.
The QES job-postings luncheon has a partner-side corroboration in RavenPack’s event page, which names Kevin Cosgrove, then Head of US Sales, as the featured speaker and describes uses of job data around strategic direction, technical-skill demand, hiring growth or decline, and geographic expansion. A public post by RavenPack CEO Armando Gonzalez further describes Wolfe’s Jobs, Skills, and Locations (“JSL”) model, built from job counts, skills in descriptions, and geography. The post’s backtest figures are vendor-reported and unvalidated here; they are a replication lead, not a published result.
The public “Seven Sins of Investing with GenAI” webcast page remains title-only and gated. A public post by Martin Brueckner places a 2026 Yin Luo report in a lineage of earlier quantitative-investing “Seven Sins” reports. It is a retrieval lead only: the report’s contents, speakers, model details, and results remain unresolved.
The historical lineage is now directly recoverable through the public 2014 Deutsche Bank Quant Handbook paper, also indexed by a Rutgers seminar record. Yin Luo and collaborators enumerate survivorship, look-ahead, storytelling, data-snooping, turnover/cost, outlier, and shorting/asymmetric-pattern pitfalls. This is historical validation methodology and personnel lineage; it does not reveal the 2026 Wolfe GenAI report or any current model deployment.
The EventSmart Yin Luo biography adds a public personnel route: it describes Luo as Wolfe’s Vice Chairman and QES leader, with previous quantitative-strategy leadership at Deutsche Bank and degrees from Renmin University, the University of Windsor, and the University of Toronto. This is biography evidence only; it does not identify the staffing or permissions behind every model.
The full capture note is here.
Balyasny Boston and Stanford routes — August 19, 2026
The title-blind speaker search recovered two university-hosted Balyasny routes that were not in the earlier conference capture. The MIT CSAIL Alliances event page records a March 4, 2026 Cambridge talk on the public state of AI in investing and names Charlie Flanagan (Chief AI Officer), Dima Tymofieiev (Senior Research Engineer), and Peter Anderson (Head of Applied AI Research). The page connects the event to machine learning, NLP, systems engineering, quantitative finance, and real-world decision-making. It is event metadata; no recording or slides were recovered.
The Stanford ICME Lunch & Learn page records a January 30, 2025 GenAI talk with Flanagan, live demos, and a separate career account by Evelyn Wang, identified there as a Senior Data Scientist. The page says the event was livestreamed, but a public replay was not located. The ACL Anthology author/research surface also links Peter Anderson and Charlie Flanagan to a finance-domain embedding paper. That paper is retained as an academic artifact, not as proof of a current production model or trading signal.
The full personnel and capture ledger is in the Balyasny media expansion note.
Future Alpha official speaker-profile expansion — August 19, 2026
The official Future Alpha 2026 programme and organizer biographies add a personnel layer beyond the secondary agenda and post-event coverage:
| Person / organization | Public profile signal | Boundary |
|---|---|---|
| Simon Chao — Fidelity Investments | Head of Asset Management Emerging Technologies; organizer text describes a global team of data scientists and full-stack engineers and names NLP, ML, and deep learning. | Organizer biography; no Fidelity model inventory, permission map, evaluation, or result evidence. |
| Tom Taylor — Man Group | Co-Head of Front Office Engineering; the biography connects the role to Man’s investment engines and Man Numeric/Man GPM research and trading systems and lists University of Warwick training. | Dated organizer profile; no direct AI-system disclosure or recording. |
| Ioana Boier — NVIDIA | Senior Principal Applied Researcher and Solutions Architect; the profile names prior Alphadyne and Citadel experience, Purdue computer-science training, publications, and patents. | Vendor biography and historical lineage; prior-firm associations do not establish deployment. |
| Paul Ebner — One River Digital | Portfolio Manager and Researcher; the biography names prior CPP Investments and BlackRock systematic-investing roles. | Dated biography; no current or historical model inventory, permissions, or results. |
| Hamza Bahaji — Amundi | Head of Financial Engineering and Investment Solutions; profile names prior Seeyond/Natixis quantitative-research leadership and Paris Dauphine doctoral training. | Organizer biography and career lineage only. |
These pages are useful routes for future speaker-post, paper, and recording recovery. The organizer’s post-event highlights link is not yet a durable recording or transcript in the reviewed public surfaces. The expanded capture note is here.
CFA Netherlands speaker-to-media cross-link — August 19, 2026
The named CFA Netherlands speaker programme also opened a title-blind media route for Harold de Boer of Transtrend. The Top Traders Unplugged interview has a dated transcript-like publisher page and timestamps around whether ML/AI language corresponds to robust process, the risks of unsupervised black-box strategies, and model robustness. A separate The Derivative episode provides a newer guest-media route, while Transtrend’s first-party role page anchors the person and firm association.
This is a speaker and methodology cross-link, not a current Transtrend AI system disclosure. The full record is in the coverage ledger.
Post-event coverage recovered — Future Alpha 2026
The initial Future Alpha record was agenda-only. A follow-up publisher search found three Hedge Fund Alpha post-event pages published April 6–8, 2026. They are secondary editorial coverage, not recordings or verbatim transcripts; the accessible portions are retained with that boundary in the capture note.
A second follow-up recovered Future Alpha’s official event-floor LinkedIn post and a speaker-network post by Florian Ielpo. Together they add first-party/event-network evidence for the Bryan Kelly keynote framing—ML and market risk, the relationship between econometrics and AI, and implications for systematic investing—and expose a vendor/sponsor graph for a separate media search. They remain event-post metadata, not recordings, slides, model disclosures, or performance evidence.
A further title-blind pass found four sponsor interviews with visible LinkedIn transcripts: YellowDog’s Bruce Beckloff, Clearwater Analytics’ Zoubair Esseghaier, Neuralk AI’s Jeremy Ben Sadoun, and OptionMetrics’ William Ko. They add vendor-reported vocabulary around data normalization, risk and portfolio infrastructure, AI agents, tabular foundation models, and research workflows. These are event-marketing and vendor-media sources: none identifies a hedge-fund customer or establishes adoption, permissions, production deployment, or investment results.
A title-blind follow-up of the same sponsor graph recovered Future Alpha’s interview with Billy Connors of SpiderRock. The visible transcript discusses integrating AI into workflows while retaining human judgment, data readiness and interpretability, and the shift from model experimentation toward trusted market infrastructure. It also points to SpiderRock’s European expansion and public descriptions of REST/WebSocket delivery plus historical intraday volatility-surface, quote/print, and close-mark datasets. The post is a vendor interview and its automatic transcript contains recognition errors; it does not establish a hedge-fund customer, model deployment, or performance result.
SpiderRock’s own news archive exposes a related data-and-AI partner graph: CPZAI describes live and historical options analytics, implied-volatility surfaces, Greeks, and historical options datasets inside an AI-native systematic research, back-testing, execution, and governance environment; QuantCrunch describes live options analytics alongside theoretical pricing, strategy back-testing, and risk workflows; and BMLL describes SpiderRock options print data in the BMLL Data Lab and a joint white paper on cross-asset options/equity research, dealer-gamma positioning, hedging flows, and intraday price behaviour. These are first-party partnership announcements and vendor product claims. They show a discoverable data/workflow surface, not a named fund implementation or measured alpha.
| Surface | Public evidence | Capture boundary |
|---|---|---|
| AI and investment-process panel | Names Wangshu Yang of Goldman Sachs QIS Alternatives, Oliver Faltin-Trager of Wellington fixed-income research, Rahul Gupta of DRW quantitative research and trading, and moderator Charles Roberts of ARK Invest. The accessible text frames current AI use around workflow tasks such as research writing, data cleaning, idea sorting, and broader coverage. | Secondary editorial summary; no recording, slides, model inventory, permissions, evaluation protocol, or return attribution. |
| AQR / Bryan Kelly | Identifies Kelly’s opening presentation and exposes topic framing around model complexity, approximation, variance, double descent, S&P 500 forecasting, and cross-sectional asset pricing. | Secondary coverage; exact slides, data, implementation, production use, and results remain unverified. |
| CSCB Management / Jacob Amaral | Identifies Amaral as Head of Quantitative Research and reports a comparison using 140 proprietary trading systems across 11 markets and seven usable sectors, with daily P&L mapped by market, sector, bar size, and trading hours. | Secondary report of firm/presentation material; the sample construction, scaling, and results require the original presentation or direct firm confirmation. |
Additional observed routes
| Lane | Evidence | Follow-up |
|---|---|---|
| Anthropic Code /w Claude | Man Group Head of Data and AI Tushara Fernando publicly described speaking at the London event about how Man applies AI | Track the next city and recover any public recording or practitioner materials. |
| Milken / fiduciary investor rooms | Citadel CTO Umesh Subramanian spoke at Milken on AI in finance; Bridgewater’s Nina Lozinski appeared at the 2025 Stanford Fiduciary Investors Symposium on AI and investment strategy | Track future programmes, speaker pages, and recordings; access is likely nomination- or invitation-led. |
| Bloomberg Enterprise Data & Tech Summit | Citadel’s CTO appeared in a Bloomberg conversation on AI, data, and investment technology | Track editorial, invitation, customer-reference, and recording routes. |
| J.P. Morgan Quantitative Investment Solutions Forum | WorldQuant’s Paul Griffin and Cubist’s Pusheng Zhang appeared on an AI execution panel | Monitor future editions and seek public programme, speaker, or partner evidence. |
Academic and research routes in the evidence set
Ramit Sawhney’s public history shows a different pattern from executive finance speakers: ACL, EMNLP, NAACL, AAAI, IJCAI, UAI, WSDM, WWW, ICASSP, AISTATS, Interspeech, and related venues. Those are worthwhile when the proposal is a real research contribution, not when the objective is general networking.
The following paper and workshop lanes are represented in the current evidence:
- FinNLP and finance-NLP workshops: strong topical fit, but the current 2026 direct deadline has passed; target the next cycle and ARR workshops.
- AI in finance academic workshops attached to major ML/NLP venues: use the personnel graph to identify authors from Tower, Bridgewater, WorldQuant, and finance-data vendors.
- Georgia Tech QCF / Financial Services Innovation Lab: local relationship route for lectures, practitioner panels, and research talks.
- NYU Stern fintech and finance-AI programs: Gideon Mann’s prior NYU Stern Fintech Conference appearance shows that the room has previously included a finance-AI leader; monitor the next dated edition rather than treating the 2024 event as current.
- ICML, NeurIPS, ICLR, AAAI, and MLSys workshops: apply only with a publishable evaluation, causal reasoning, retrieval, agent reliability, financial NLP, or efficient-model contribution. Do not travel solely for a main-conference badge.
Speaker-to-room evidence
| Person | Public speaker trail | What it tells us |
|---|---|---|
| Charlie Flanagan, Balyasny | MFA Network Boston; Artefact AI for Financial Services NYC; CQF practitioner profile | MFA and specialist finance-AI rooms can expose a named applied-AI owner. Boston MFA programming is a local route. |
| Ramit Sawhney, Tower | Georgia Tech AI and Future of Finance; AI-for-finance teaching; ACL/EMNLP/AAAI/UAI and other research venues | A technical finance-AI speaker can bridge hedge-fund rooms, university programs, and top academic venues. |
| Tushara Fernando, Man Group | Anthropic Code /w Claude in London; current Man Group profile confirms firmwide data and generative-AI remit | Vendor engineering conferences can surface real buy-side implementation details when the speaker is an AI owner. |
| Gideon Mann, Millennium/Bloomberg | NYU Stern Fintech Conference; Stanford AI in Fintech Forum; EMNLP invited talk; D4GX | NYU, Stanford fintech, data-for-good, and NLP communities are useful person-discovery surfaces, even when the original event is past. |
| Umesh Subramanian, Citadel | Milken Global Conference; Bloomberg Enterprise Data & Tech Summit; Reuters NEXT | Executive technology rooms are relationship-led and valuable for competitive intelligence, but usually poor cold-CFP targets. |
| Nina Lozinski and Oliver Simon, Bridgewater | Stanford Fiduciary Investors Symposium; Aspen Technology Leaders Initiative; Bridgewater AIA public research | Fiduciary, institutional-investor, and technology-leadership programs are targeted routes for AIA-level public access. |
| Paul Griffin, WorldQuant | J.P. Morgan Quantitative Investment Solutions Forum; ICML sponsorship and AI/data programming | Quantitative-investment forums and major ML sponsorship ecosystems are the right routes for WorldQuant and comparable firms. |
Application checks
For each event, record whether the public evidence establishes:
- a named hedge-fund, prop-trading, or asset-manager practitioner;
- a speaker, CFP, moderator, attendee, or nomination route;
- a relevant investment, research, data, risk, or AI-platform audience;
- a location or relationship context;
- a public-safe topic and a concrete capture or meeting deliverable.
An absent field remains an open question. It is not converted into a negative comparison or a comparative judgement.
Reusable public topic package
Prepare one reusable, non-confidential talk with three variants:
- Buy-side operator: “From Copilots to Controlled Agents in Investment Research”
- Technical practitioner: “Evaluating Retrieval, Tool Use, and Human Review in Financial Agents”
- Executive audience: “The Operating Model for a First AI Hire at an Investment Firm”
Each version should use public examples, explicit evaluation methods, and no fund-specific architecture, holdings, or performance claims.
August 27 personnel and future-conference expansion
The official Johns Hopkins AMS 2026 agenda lists Michael Baeder of Campbell and Company and Chaoyu Liu of Singular Square Management as keynote speakers on April 7, 2026. The organizer’s speaker biographies describe Baeder’s global-equity, data-analysis, and machine-learning remit, plus a biweekly research seminar, AI working group, and internal research peer review. The same page describes Liu’s AI-and-alternative-data research platform, with interests in market microstructure and real-time information processing. The event page is a personnel and topic route; no recording or talk transcript was found.
Baeder’s public titles conflict across surfaces: the Johns Hopkins organizer says “director,” while The Org lists Principal Researcher. Campbell’s own site independently describes proprietary data, high-speed infrastructure, contextual risk management, and a human/statistical/infrastructure process, but does not resolve the title or expose a specific AI system. The discrepancy is retained for later verification.
The Jacobs Levy Center 2026 conference agenda adds a future September 25, 2026 Wharton route. It lists “Machine Learning Meets Markowitz,” Campbell Harvey and Winston Wei Dou, State Street’s Jennifer Bender as a discussant on traded risk factors, and a fireside chat with AQR’s Cliff Asness. The agenda links papers and provides programme metadata; it does not establish recording availability, shared technology, deployment, permissions, or results.
See the Johns Hopkins capture note for source-level boundaries.
August 31 QuantVision and Cornell Financial Engineering Manhattan routes
The public QuantVision 2026 programme adds a past March 19–20 Fordham event whose agenda uses transformer, graph-model, deep-learning, and regime-robustness vocabulary. The page lists Samson Qian and Arkin Gupta from Citadel, Matt Rowe from Man Group, Mike Tiano from Schonfeld, Claudia Perlich from Two Sigma, and data-sourcing participants connected with Millennium, T. Rowe Price, RavenPack, and other organizations. It states Chatham House rules; no recording or session transcript was located. The agenda therefore functions as a speaker and topic-discovery route only.
The Cornell Financial Engineering Manhattan 2026 programme, published August 28 for September 11, lists panels involving Citadel, Millennium, Cubist/Point72, Magnetar, BlackRock, Wolfe Research, Tower Research, Two Sigma, CFM, T. Rowe Price, and others. Its session labels include LLM look-ahead bias, temporal information constraints, data leakage, validity of LLM-based financial prediction, future data for modeling, and AI in investing and trading. These labels identify the public agenda, not delivered remarks or firm implementations. Attendance, current affiliation, model ownership, permissions, and performance remain unverified until first-party or event materials are recovered.
The conference capture note records the promoter-source boundary and the follow-up queue for official recordings, slides, speaker posts, and post-event materials.
September 2 STAC regional calendar expansion
The STAC events calendar adds a concrete follow-up sequence for the trading-technology and machine-learning audience: Sydney on September 10, Tokyo in English and Japanese on September 15, London on October 6, and New York on October 21, 2026. The Tokyo roster includes personnel from Rakuten Institute of Technology, JPX, Morgan Stanley, Mizuho, and Pico alongside infrastructure vendors. The London roster includes BestX, Calvin Risk, TimeStored.com, AWS, and INQDATA. The New York roster includes Barclays, INQDATA, AWS, Expanse, JetCool, and Penguin Computing.
These are advertised event dates and displayed speaker affiliations. They do not establish attendance, internal AI projects, model or dataset use, agent permissions, or investment results. The agenda is partly embedded, so the capture queue should resolve the agenda payload and check each event after it occurs for lawful recordings, slides, transcripts, and speaker follow-up. See the regional STAC source note.
Sources checked
- Tower: AI in Capital Markets
- Man Group: Tushara Fernando
- Tushara Fernando on Code /w Claude
- NYU Stern 2024 Fintech Conference
- MFA Network Boston speakers
- Milken: Umesh Subramanian
- Reuters NEXT New York speakers
- Citadel Enterprise Data + Tech Summit
- Bridgewater: Nina Lozinski
- WorldQuant: Paul Griffin
- AI Loves Data New York
- CFA Society Netherlands: AI, Machine Learning and Agentic Investing
- MathFinance Conference 2026
- CEPR Imperial–Goldman Sachs Hedge Fund Conference
- QuantMinds Quant Invest Summit
- Future Alpha 2026 agenda
- Future Alpha 2026 AI/investment-process coverage
- Future Alpha 2026 AQR/Bryan Kelly coverage
- Future Alpha 2026 CSCB/Jacob Amaral coverage
- CQF Institute AI and Machine Learning in Quant Finance Conference
September 1 title-blind conference follow-up
The latest title-blind pass added seven event routes that are useful for monitoring but should not be mistaken for firm implementation evidence: J.P. Morgan’s 2026 QIS conference summary, the With Intelligence Hedge Fund COO Summit Europe agenda notice, the Fall JOIM AI in Finance announcement, the Cornell Future of Finance & AI route, the ICCF 2026 Oxford announcement, the AQFC 2026 programme book, and the London AI in Financial Services speaker preview. Their public material exposes dates, topics, organizer framing, or advertised roles. It does not establish attendance, session content, model ownership, private datasets, permissions, production deployment, or investment results. The detailed capture and append-only ledger are in the source note and the conference addendum.
Source status: Public event and programme metadata checked on 2026-09-01; session recordings and transcripts were not recovered in this pass.
Source files: sources/13-multimodal-sources/conference-title-blind-followup-2026-09-01-raw.md; sources/13-multimodal-sources/conference-title-blind-followup-2026-09-01-addendum.json