Our personnel coverage is strongest where a firm publishes both a named AI remit and a workflow or platform description. It is weaker where the public record contains only a job family, a generic technology title, or a historical ML profile. That is an evidence-coverage distinction, not a judgment about the people or private capability.
The existing roster should therefore be read in layers:
- Named AI owner: current person and explicit AI/GenAI remit.
- Named AI practitioner: person connected to a disclosed workflow, lab, or artifact.
- Role-led AI function: current job description exposes architecture or investment use, but the person is not public.
- AI-adjacent investment/technology leader: CTO, CIO, quant, or research title connected to AI by a separate source.
- ML-only or discovery signal: machine-learning or quantitative role without public GenAI ownership.
Personnel evidence coverage by firm
| Firm | Current public personnel coverage | Strongest signal found | Main unresolved gap |
|---|---|---|---|
| GMO | AI-adjacent investment and technology leadership | Hylton Socher is CTO/partner with ML portfolio background; GMO also exposes NLP/alpha infrastructure | No current named GenAI owner or public platform roster |
| Acadian | Named/role-led investment-AI coverage, with qualification | VP Investment AI Engineer role; management describes agentic coding and AI investment; Javier Alcazar supplies ML/NLP investment background | VP identity, reporting line, and production status are not firm-roster confirmed |
| Arrowstreet | Role-led platform coverage plus named discovery leads | Senior AI Platform Engineer specifies Bedrock, routing, MCP, RAG, agents, telemetry, and controls; Florent Monthel publicly recruits and publishes a safety artifact | Formal personnel title and firm deployment are not independently established |
| CFM | Named research-side AI coverage | Eric Vanden-Eijden is publicly tied to an internal AI effort; the ML Lab and Philip Seager interview expose agentic strategy research | Reporting lines, model inventory, and current GenAI platform ownership remain unclear |
| Balyasny | Named central AI and practitioner coverage | Charlie Flanagan is identified as Chief AI Officer; Applied AI, data, and engineering personnel are also public | Complete roster, model ownership, and decision permissions are not public |
| Man Group / AHL | Named firmwide AI plus research-workflow coverage | Tushara Fernando is Head of Data and AI; AlphaGPT/AlphaTrend authors and central-platform personnel are public | Day-to-day ownership of specific systems and live capital attribution are not public |
| Bridgewater | Named lab, investment, applied-AI, and technical coverage | Oliver Simon/Nina Lozinski have AI & ML Investment Strategy titles; PAT presenters identify applied-AI, investor, and technical leads | Current team structure and system-by-system authority are incomplete |
| Jane Street | Strong ML trading personnel context; weak named GenAI ownership | Firm publishes deep-learning trading infrastructure and internal AI-assistant work | No public firmwide GenAI lead or direct AI-assistant/trading link |
| QRT | Role-led AI-platform coverage; named quant/AI-adjacent researchers | AI Platform Engineer, AI Delivery Engineer, and AI/LLM security roles expose platform and control vocabulary | No named current AI platform owner or deployment evidence |
| Tower Research Capital | Named direct AI/ML leadership | Ramit Sawhney is identified as Global Head of Core AI & ML; firm also discusses agents and research | Model ownership, production boundary, and performance attribution are not public |
| WorldQuant | Named AI-leadership coverage plus role-led agentic PM signal | Paul Griffin is identified as leading AI initiatives; current AI/LLM careers and dated agentic-PM role | Named assignment to the agentic portfolio-manager system is not public |
| Numerai | Platform and role-led coverage | AI Scientist role, Predictive LLM, open skills, and MCP tournament tooling | No named internal AI owner or autonomous capital authority evidence |
| Point72 / Cubist | Role-led GenAI, ML, security, and investment-AI coverage | Current roles mention GenAI technology, MCP agents, synthetic data, production-support agents, and AI security | No public firmwide GenAI owner or filled-role map |
| Schonfeld | Named/role-led investment enablement coverage | Official AI Lab and leadership pages; public materials identify AI initiatives and internal tools | Reporting lines, model routing, and complete platform-team roster are not public |
| G-Research | Role-led infrastructure and research coverage | Core AI, Applied AI, NLP, MCP, on-prem inference, fine-tuning, and secure sandbox roles | No public named AI executive or investment-workflow owner |
| AQR | Named ML/research leadership, limited GenAI coverage | Bryan Kelly and public ML/NLP research establish method and research depth | No current named GenAI owner or internal agent platform found |
| Two Sigma | Strong AI-system and research hiring signal, limited named ownership | Official careers page explicitly spans AI systems; public research describes LLM-assisted feature workflows | Public source set does not establish a current named GenAI owner |
| Citadel | Named CTO/workflow coverage plus role-led systematic ML | Umesh Subramanian is tied to the Citadel AI Assistant; GQS roles mention LLMs, pretraining, fine-tuning, and RL | Assistant ownership, architecture, and GenAI-to-trading boundary remain undisclosed |
| Millennium | Direct named AI-owner coverage | Gideon Mann is Global Head of Artificial Intelligence, Technology and leads an AI advisory group | Model inventory and investment authority are not public |
| D. E. Shaw | Role-led applied-AI coverage, no named investment AI owner | Applied AI Engineer, AI product, AI vendor-tools, and fundamental-equities AI product roles | Entity/function boundaries and current leadership are unresolved |
| PDT | ML and research-engineering coverage, not GenAI-specific | Applied ML Scientist and research engineering for large-model training/fine-tuning infrastructure | No public LLM/agent owner or GenAI investment workflow |
| XTX Markets | Named predictive-ML leadership, not GenAI-specific | Atlas Wang and XTY Labs expose deep-learning/market-data research | No public LLM or agent program |
| Aspect | Named ML-aware investment leadership, not GenAI-specific | Martin Lueck and Bas Monsewije discuss ML/AI/LLM research constraints | No named GenAI platform owner or deployment evidence |
| Voleon | Named predictive-ML and technical leadership | Management page identifies CIO, CTO, Chief of Technical Staff, and predictive-model builders | No public GenAI owner or agentic investment workflow |
| Winton | Investment/data/technology coverage, limited direct AI personnel | Public investment and technology leadership discusses AI/ML context | No named current GenAI owner located |
| Systematica | Investment and data-R&D coverage, limited direct AI personnel | Leda Braga and Data R&D/software signals | No named current GenAI owner or public LLM system |
| Marshall Wace | Technology and coding-agent hiring signal, no named owner | Technology program and RLM/RAG public artifacts | No current named AI/ML leader corroborated |
| Brevan Howard | Qualified role-led GenAI coverage | Quantitative Strategist listing describes LLM/RAG/embeddings/local models; Tim Mace is a qualified Head of AI lead | Head-of-AI title needs firm-controlled confirmation |
| Caxton | Disqualified current-personnel signal | Stale/third-party Python LLM Engineer listing | No current firm-controlled AI personnel evidence |
| Squarepoint | Quantitative-ML coverage, no current GenAI personnel | Systematic research and automated strategies | No named current GenAI owner or public workflow |
| H2O Asset Management | Direct named GenAI remit | Timothée Consigny is CTO and Head of Innovation in Generative AI | Asset-manager comparison; implementation and performance remain undisclosed |
| SummitTX Capital | Named CTO with explicit AI remit | John Timotheou’s firm announcement assigns AI to the CTO portfolio | No specific system, personnel depth, or performance evidence |
What titles actually predict
The most useful non-obvious title signals are not generic seniority titles. They are combinations of title and mandate:
| Title or phrase | What it can indicate | Required corroboration |
|---|---|---|
| CTO | Technology ownership that may include AI | A source must explicitly assign AI, GenAI, agent, or model responsibilities |
| Chief Science Officer | Research/science leadership, sometimes AI ownership | Confirm AI initiative language and whether it covers the investment business |
| Head of Data / Head of Data and AI | Data platform or firmwide AI enablement | Look for model, agent, skills, or investment-workflow remit |
| Head of Research / Head of Quantitative Research | Investment research authority | AI-specific publications, roles, code, or agent workflow needed |
| Head of Platform / Platform Engineering | Runtime, data, or deployment control | Look for LLM serving, MCP, RAG, evaluation, telemetry, or permissions |
| Chief Data Officer | Data governance and source infrastructure | Do not infer AI ownership without an explicit AI or model mandate |
| Head of Investment Strategy / PM | Investment decision proximity | AI must be connected to the person’s remit; title alone is insufficient |
| AI Platform / Applied AI / AI Lab | Direct technical ownership or enablement | Confirm current affiliation, team scope, and whether the source is firm-controlled |
| AI Security / Model Risk / Agent Evaluation | Control-plane ownership | Identify the systems, business boundary, and whether the role is filled |
| Quantitative Developer / Research Engineer | Implementation proximity | Require AI/LLM/fine-tuning/agent language; generic quant development is not enough |
Search protocol for the next sweep
Search each firm’s official careers, leadership, team, research, GitHub, podcast, conference, and LinkedIn surfaces for the title families above. Run the loop in this order:
- Find title and firm-affiliation candidates.
- Verify current affiliation on a firm-controlled page or a current role.
- Find a second source connecting the person or role to AI/ML/GenAI work.
- Classify the signal as named owner, named practitioner, role-led, adjacent, ML-only, historical, or disqualified.
- Record the missing link: reporting line, filled status, system ownership, deployment, permissions, evaluation, or investment authority.
The first expansion targets are Millennium, H2O, SummitTX, Brevan Howard, Point72/Cubist, D. E. Shaw, Tower, G-Research, and firms currently represented only by ML/research titles: Winton, Systematica, Marshall Wace, Aspect, Voleon, PDT, and Squarepoint.
Second sweep: new title families and disqualification outcomes
The next search widened the vocabulary beyond “Head of AI” and found several useful patterns:
| Discovery | Signal | Disposition |
|---|---|---|
| State Street Investment Management | “Global Head of Insight Generation and AI Adoption” role covering GenAI, agentic AI, modeling, and enterprise AI platforms | Adjacent asset-manager control case; not a hedge-fund personnel record |
| Marathon Asset Management | “Head of Data Management & AI” role owns enterprise data/AI strategy, investment-research synthesis, PM partnership, AI architecture, monitoring, and governance | Role-led credit-manager signal; person not public in the listing |
| Systematic Strategies | Jonathan Kinlay publicly identifies as Head of Quantitative Trading and publishes an agentic alpha-research case study | Named practitioner signal; self-published workflow, not proof of firm-wide deployment or performance |
| ExodusPoint | Firm-owned LinkedIn post announced Shen Xu as Head of Artificial Intelligence and described AI-infrastructure work | Historical named title; current firm roster does not corroborate continued affiliation |
| Magnetar | Third-party LinkedIn material calls Trevor Mottl “Head of AI Quant” | Validation queue only; official leadership page does not show that title or person |
| Move37 Capital | Official site calls the manager AI-driven and names Kunal Gautam as Partner | Emerging-manager discovery signal; no named AI owner or independently audited AI evidence |
| Private investment-manager searches | Recruiter listings use “Head of Data & AI,” “Head of AI,” and “Head of Platform Engineering (AI)” for investment-research platforms | Search vocabulary and market-demand signals; recruiter copy is not firm-controlled evidence |
Fresh verification pass
The CFM record was rechecked against current firm-controlled pages. CFM’s current approach page describes the integration of machine learning, AI, and cloud computing with its research and data platform, while its January 2026 interview says the firm established a machine-learning lab headed by domain experts and is investing in generative AI for sentiment/context extraction, classification, risk management, automated research, and coding productivity. This strengthens the firm-level program record. It still does not establish a complete personnel roster, individual reporting lines, model ownership, live permissions, or investment performance attribution. The source ledger records these as firm-controlled strategy evidence, separate from named-person evidence.
These results change the search heuristic. A role can carry AI ownership through its object—insight generation, data management, platform engineering, AI quant, or research engineering—even when the title is not “Chief AI Officer.” The object and responsibilities must still be verified before personnel promotion.
The current AI-native quant-researcher posting makes the intended technical surface more specific: the role spans target definitions, model architectures, training recipes, distributed multi-GPU foundation-model training, evaluation, and collaboration with the ML-platform and portfolio teams. Its requirements mention mixture-of-experts and long-context transformers, vision-language models, efficient attention, multi-token prediction, distributed training frameworks, and optional LLM inference optimization, RLHF/RLAIF/DPO/reward modeling, Ray/Kubernetes, and CUDA. This is a first-party hiring specification, not evidence that each technique is implemented, that the role was filled, or that a model has live capital authority.
The personnel map can now name a second CFM research-side signal without inferring a private roster: CFM’s public ML Lab announcement names Eric Vanden-Eijnden and Anastasia Borovykh and identifies Giulio Biroli as leader of the CFM-ENS Data Science chair. Borovykh’s public profile describes her current CFM work on machine-learning sources of alpha and links a pre-CFM peer-reviewed paper on dilated convolutional networks for financial time-series forecasting. The CFM seminar post adds a dated “Generative Models for Quant Finance” presentation by Vanden-Eijnden. These are role, lineage, and topic signals; they do not establish current model ownership, production use, or returns.
Source files and evidence boundaries
Source files: sources/06-industry-verticals/hedge-fund-ai-personnel-title-sweep-2026-raw.md, research/06-industry-verticals/gmo-acadian-arrowstreet-ai-public-signals-2026.md, and sources/06-industry-verticals/gmo-acadian-arrowstreet-ai-public-signals-2026-raw.md.
The sources establish public titles, role descriptions, firm statements, and disclosed workflows. They do not establish a complete roster, current reporting lines, role fulfillment, model ownership, production state, investment authority, or performance. “No current public evidence found” is a search boundary, not a claim about private capability.
Third sweep: ownership beyond the “Head of AI” title
The next pass found four additional ownership layers that should be tracked separately:
| Layer | New public signal | What it establishes | What it does not establish |
|---|---|---|---|
| AI product ownership | Point72’s current AI Product Analyst, Market Intelligence role owns AI products from approval through testing, deployment, support, adoption, and model governance | Investment-facing AI product responsibility and explicit lifecycle language | The posting does not identify the hire or prove the products are live |
| AI validation and release control | Point72’s current AI Validation Engineer, Macro Technology role owns validation strategy, regression packs, release gates, and human review of AI-generated code/configuration | AI control-plane and trading-workflow validation mandate | It does not disclose model architecture, model ownership, or trading permissions |
| GenAI platform and infrastructure | Point72’s current careers surface lists Machine Learning Infrastructure Engineer, GenAI Technology; Full-stack Engineer, GenAI Technology; NLP/AI Engineer; and related roles | A current GenAI technology job family and platform-building signal | It does not provide a named leader or filled-role map |
| Model research and training substrate | Citadel GQS postings explicitly mention large language models, pre-training, fine-tuning, reinforcement learning, distributed training, inference optimization, and internal ML libraries | A firm-controlled research and engineering vocabulary connecting ML infrastructure to systematic-investing workflows | It does not establish a particular production model, strategy, or return attribution |
Additional geographic and emerging-manager signals:
| Firm | Public signal | Disposition |
|---|---|---|
| High-Flyer | Official English site describes AI-focused quantitative trading, neural-network and NLP research, a first deep-learning trade in 2016, and the internally built Fire-Flyer deep-learning platform | Firm-controlled China-based AI-quant disclosure; self-reported and not an independent technical or performance audit |
| Ubiquant | Official quantitative-competition page, Ubiquant-linked Hugging Face organization, and a dated public recruiting post | Public AI-research artifact surface; competition/talent funnel; roles named Quantitative Strategy Researcher, AI Algorithm Researcher, Data Scientist, Quantitative Implementation Engineer, and Quant Developer |
| M37 Management LP | University-hosted AI Engineer posting describes agents, Claude Code, MCP, multi-agent collaboration, benchmarks, security, and integration with investment research; SEC and LinkedIn records corroborate the manager’s existence | Emerging-manager role-led signal, with secondary job-post evidence; distinct from Move37 Capital and not a firm-controlled posting |
| Trexquant | Current careers page describes machine learning for alpha discovery and portfolio construction, with global alpha-research roles | ML/alpha research signal; no current public GenAI or agent ownership found |
The resulting search rule is broader: search for the person or team that owns the AI product lifecycle, validation evidence, model infrastructure, or investment-workflow integration—not only the executive whose title contains AI.
Fourth sweep: “Head of Automation” and its neighboring titles
“Head of Automation” should be added to the search vocabulary, but it cannot be promoted as an AI title without examining the object being automated:
| Title signal | Public example | Classification |
|---|---|---|
| Risk Automation Lead | Man Group’s current AHL role drives automation and AI adoption across investment-risk teams, builds Python risk tools, and integrates LLMs into workflows | Direct investment-control and AI-adoption signal; role-led, person not identified |
| Security Automation Lead | Point72’s current role owns a single auditable security-automation pipeline in a technology organization that uses AI solutions | Control-plane automation signal; not an investment-research AI owner |
| Head of Automation, Analytics and Platform Services | State Street’s Nick Delikaris role spans algorithmic trading, business-process engineering, analytics, and platform services | Trading and platform automation signal; adjacent asset-manager evidence |
| Global Head of Automation | MarketAxess’s Gareth Coltman role concerns automated fixed-income execution and Auto-X | Market-infrastructure and execution-automation evidence; not hedge-fund personnel |
| Head of Automation and AI | Alter Domus’s Davendra Patel role covers proprietary AI systems and fund-administration workflow automation | Fund-services control case; not a hedge-fund investment-team record |
| Head of Development, Automation and AI | Recruiter listing for an investment-management house combines development leadership, automation, and AI/ML integration | Discovery vocabulary only; third-party copy and unnamed firm |
The important inference is organizational: “automation” may be the operating bridge between a research or risk team and an AI platform. It often reveals workflow ownership, release controls, and adoption responsibility even when the word GenAI is absent. It does not, by itself, establish language-model use, autonomous investment decisions, or model ownership.
Evidence Boundaries
The coverage labels describe the strength and completeness of the public personnel trail. They do not describe the quality of any person, firm, strategy, model, or AI system.