Head of Quantitative Research

Protogon Research•San Diego, CA

11d•Onsite

About The Position

Protogon Research is an elite team based in San Diego, CA, backed by top VCs including OVO Fund, Zelda Ventures, and others. Led by serial entrepreneur Rafael Cosman, co-founder of Archblock and the TrueFi DeFi protocol, we design autonomous AI systems and deploy them directly into financial markets through proprietary trading. This is where research meets the real world, and where models must perform, adapt, and improve continuously. Financial markets provide clear feedback and real consequences, which gives us several key advantages: Clear feedback: progress is measurable, and weaknesses surface immediately, forcing our systems to move beyond simple pattern recognition or emulating human behavior toward deeper understanding and super-human performance. Long-term focus: unlike traditional quant funds, we're optimizing for the long-term success of our AI technology, not short-term profits. Original work: unlike much of the AI industry, we're not layering a thin wrapper over existing models. Our work is original, proprietary, and defensible. We're seeking an executive-level Head of Quantitative Research to own Protogon's trading P&L and lead the elite technical organization that produces it: model development, research direction, risk, capital allocation, and execution. You will lead our AI research direction, as well as how our models connect to markets to generate revenue. We trade digital assets today and intend to expand into additional instruments and markets, with this role leading that expansion. Crypto experience is welcome but not required. You will report to the CEO and lead our technical organization, including our AI/ML and quant dev team. You will be deeply hands-on from day one and will thoughtfully lead and grow your team. We work primarily in person and are looking for candidates based in or willing to relocate to the San Diego area.

Requirements

Roughly 8-12 years in quantitative research or applied ML from a systematic fund, proprietary trading firm, or crypto market maker.
Models have carried real capital and you answered for the outcome.
Made hard calls under pressure about whether an underperforming strategy was broken or simply in a bad regime.
Personally designed models and experiments rather than only deploying existing ones.
Fluent in Python.
Comfortable with modern ML techniques such as deep learning, sequence models, reinforcement learning, and time-series modeling.
Know what separates a backtest artifact from a durable edge: leakage, overfitting, regime dependence, unrealistic fill assumptions.
Understand that realized P&L is a function of costs, capacity, venue economics, and sizing, not just signal quality.
Can look at a live result and tell whether the model, the execution, or the market changed.
Led research or technical work and want to stay close to it rather than drift into pure management.
Comfortable navigating considerable ambiguity, and building the process yourself.
Want to build AI systems in an environment that tests them against real outcomes, and you can hold near-term P&L accountability alongside longer-term research goals.
Current authorization to work in the United States; Protogon Research is not able to provide visa sponsorship at this time.

Nice To Haves

PhD in a quantitative discipline, or equivalent research depth
Depth in market microstructure and algorithmic execution
Experience building a research or trading capability inside an early-stage company

Responsibilities

Own trading performance and hit the targets we set together.
Own the constraints our systems operate inside: position sizing, leverage, exposure and concentration limits, and drawdown protocols.
Build the attribution that makes performance clear, decomposing returns into model alpha, market exposure, execution quality, financing, and fees, and explaining divergence between simulation and live results with precision.
Lead our technical organization, including our machine learning and software engineers.
Set the modeling roadmap, prioritizing it against where capital is actually at work, and hold the team to the standard that live results demand.
Own hiring, onboarding, and development in partnership with company leadership, and grow the quantitative and trading capability the book will require at scale.
Own the research agenda and the models it produces, from hypothesis and experiment design through deployment into live trading.
Improve performance, robustness, and adaptability across changing market conditions, with a bias toward measurable real-world impact over novelty.
Raise the rate at which good ideas reach production.
Direct the development of core ML infrastructure, data pipelines, training workflows, evaluation tooling, and simulation fidelity, so that models are reliable, observable, and honest about what they will do in live markets.
Own realized execution quality: venue selection and routing across fragmented liquidity, fee tiers and maker/taker economics, slippage, and transaction cost analysis, alongside margin and collateral management, settlement, custody, and counterparty exposure.
Institutionalize the operating layer through monitoring, reconciliation, and tested kill switches.