Applied Research - Quantitative Finance

Prime Intellect•San Francisco, CA

6d•$150,000 - $300,000

About The Position

Prime Intellect is hiring an Applied Researcher focused on quantitative finance to build environments, evaluations, agents, and post-training methods for quantitative research and systematic decision-making. The role applies agentic reinforcement learning to financial settings with noisy, non-stationary observations, delayed and path-dependent actions, and risks from misleading objectives or historical results that fail under realistic costs, constraints, or new market regimes. Work may span research agents, market and portfolio simulators, time-series reasoning, execution, risk, and production feedback loops connecting hypotheses to deployed systems. The company has active programs with quantitative funds and other finance teams. Applied Research at Prime turns workflows into tasks, environments, evaluations, and reward signals; runs post-training experiments against real baselines; investigates whether apparent gains come from the model, data, harness, inference, or infrastructure; and deploys successful approaches. Reusable work may become shared research, open-source systems, or product capabilities. Applied Research roles are externally facing, and most researchers work directly with customers, users, or research partners while maintaining a substantive research agenda. Representative early work includes building a quantitative-research environment where agents form and test hypotheses using analysis and backtests on point-in-time data; designing rewards and evaluations that detect leakage, overfitting, unrealistic execution assumptions, concentration, and tail risk; and training long-horizon agents to use data, research, coding, simulation, and portfolio tools while preserving provenance, respecting risk constraints, and adapting to changing market conditions. The researcher will own environments incorporating realistic market mechanics and operational constraints; agents that conduct research and recover from failed hypotheses; verifiers and evaluations of robust out-of-sample behavior; and production loops with monitoring, human authority, regression gates, and failure attribution. The role offers flexible work from San Francisco, New York, London, or remotely, and lists cash compensation of $150,000–$300,000 plus equity incentives.

Requirements

Depth in reinforcement learning, agents, or post-training systems, with the ability to become fluent in quantitative research.
Hands-on experience building and debugging model-training, post-training, agent, evaluation, or closely adjacent experimental loops.
Strong experimental judgment, including meaningful baselines, careful trace inspection, clear failure attribution, and willingness to change direction when evidence changes.
Strong Python skills and the ability to build artifacts other researchers and engineers can run, inspect, and extend.
Comfort owning ambiguous problems and keeping partners and teammates current while resolving them.

Nice To Haves

Prior work in systematic trading, market microstructure, time-series modeling, econometrics, or financial simulation is useful but not required.
Evidence of exceptional work such as publications, open-source systems, benchmarks, or shipped products; applicants are encouraged to describe something difficult they built, what failed, and how evidence changed their approach.

Responsibilities

Build quantitative-research environments grounded in point-in-time data and realistic market mechanics, transaction costs, latency, capacity, and operational constraints.
Develop agents that conduct quantitative research, use analytical and execution tools, inspect evidence, and recover from failed hypotheses.
Design verifiers and evaluations that test robust out-of-sample behavior across assets, horizons, regimes, and adverse conditions.
Develop post-training methods and run experiments against meaningful real baselines.
Build production feedback loops with monitoring, human authority, regression gates, and failure attribution across models, data, simulation, and infrastructure.
Turn workflows into tasks, environments, evaluations, and reward signals, then deploy approaches that work.
Work directly with customers, users, or research partners while maintaining a substantive research agenda.
Create reusable research, open-source systems, and product capabilities that may transfer to other noisy, high-stakes decision systems.