Feedback Loop Implementation: Design and implement systems that connect trade outcomes back to strategy improvement, specifically focusing on signal selection, risk parameters, position sizing, and timing.
Evaluation Frameworks: Build frameworks to quantify which signals and market conditions accurately predict profitable trades versus noise.
Automated Strategy Generation: Develop systems to explore new configurations, backtest them against real fleet data, and surface candidates for deployment autonomously.
Market Adaptation: Build mechanisms to detect shifts in market conditions (e.g., trending vs. choppy) and adapt fleet behavior in real-time.
Fleet Monitoring: Create higher-order agents for automated monitoring to catch configuration errors and performance degradation across all concurrent agents.
Performance Attribution: Decompose trades into component drivers—signal accuracy, execution efficiency, and exit timing—to feed insights back into strategy design.
Coordination & Risk: Manage concentration risk and capital allocation across the fleet, balancing the exploration of new approaches with the exploitation of proven strategies.
Infrastructure Ownership: Transition from external LLM dependence to controlled intelligence, evaluating hosting strategies ranging from proxied external models to fine-tuned, domain-specific models.
Data Capture: Build the telemetry and data capture layer to ensure every decision and outcome is structured and queryable.
Domain-Specific Training: Determine the efficacy of domain-specific training over general-purpose prompting and build the necessary pipelines for implementation.
Inference Optimization: Optimize inference for many concurrent agents, ensuring structured decision outputs and cost-efficiency at scale.

Requirements

Production ML Engineering: Proven experience training, deploying, and maintaining models that run in production and directly impact business outcomes.
Reinforcement/Online Learning: Deep understanding of the practical challenges of learning from real-world outcomes rather than static datasets.
Closed-Loop Systems: A track record of building systems where predictions lead to actions that generate outcomes, which then feed back into improved predictions.
Software Engineering: Proficiency in Python is required, with additional comfort in Go or TypeScript for production services. Experience building data pipelines and distributed systems is essential.
Preferred Experience: Background in signal generation, alpha research, portfolio optimization, or execution.
LLM Specialization: Experience with fine-tuning and serving (PEFT/LoRA, vLLM, TGI) or custom inference pipelines.
Multi-Agent Systems: Experience designing environments where autonomous agents coordinate or learn from one another.
Domain Knowledge: Background in on-chain data, DeFi protocols, or sectors where agents make sequential decisions under uncertainty (e.g., robotics, game AI).

Benefits

Base Salary: $175,000 – $250,000 USD (dependent on location and experience).
Equity: Approximately 1% initial stock grant, with significant valuation growth potential.
Performance Incentives: Eligibility for salary increases and bonuses tied directly to revenue and usage.
Token Participation: Pro-rata participation in the client’s planned 2026 token launch.
Ownership: High-impact role with meaningful upside tied directly to the success of the autonomous fleet.

Applicant Tracking System Keywords

Tip: use these terms in your resume and cover letter to boost ATS matches.

Hard Skills & Tools

PythonGoTypeScriptMachine LearningReinforcement LearningClosed-Loop SystemsData PipelinesDistributed SystemsInference OptimizationAutomated Strategy Generation

Soft Skills

CoordinationRisk ManagementPerformance MonitoringStrategy DesignAdaptabilityProblem SolvingAnalytical ThinkingCollaborationCommunicationDecision Making