Quantitative Developer – Agentic Research Lab

Role and platform

Platform engineering and agents

• Set the agentic research architecture and ship production agents that formulate hypotheses, run experiments, validate results, and monitor strategies.
• Put strategies, signals, datasets, prompts, models, and evaluations through versioned CI/CD, with reproducible builds, validation gates, staged deployment, and rollback.
• Connect agents to live Kafka or Redpanda streams and ClickHouse; design topics, schemas, views, replay, backpressure, and appropriate delivery semantics.
• Build a multi-agent workflow for hypothesis generation, experiment specification, simulation execution, and critical review.
• Expose research capabilities through typed tools, including an MCP server over the research SDK, campaign engine, and experiment registry.
• Run scheduled research rounds on fresh data and record configuration hashes, data snapshots, and code versions.
• Monitor strategy performance and route alerts to agents for context gathering, triage, and incident-summary drafts.
• Own controlled promotion from shadow mode to paper trading and small live allocations, with criteria, kill switches, and rollback drills.
• Build experimentation frameworks and safeguards against overfitting, including trial budgets, multiple-testing correction, inaccessible holdouts, and independent re-execution.
• Evaluate agents with known-answer tasks; route models by task, track costs, and operate secure on-prem GPU inference for open-weight models.

Must-have qualifications

• 10+ years building production software or ML systems, including several years as a technical lead setting architecture.
• Production agent experience covering orchestration, tool calling, evaluation, red-teaming, model routing, costs, and human-in-the-loop rollouts.
• Hands-on quantitative experience building backtests, factor or signal models, or risk models; knowledge of look-ahead bias, data snooping, regime change, and transaction costs.
• Platform-level MLOps and CI/CD experience building registries, experiment tracking, validation, deployment pipelines, and reproducible workflows.
• Production streaming and analytical data experience with Kafka or Redpanda and ClickHouse, including schemas, partitioning, consumer lag, and query performance.
• Strong Python and SQL; Linux debugging and Kubernetes workload experience with infrastructure as code.
• Experience shipping automated decision systems in regulated or capital-at-risk environments with kill switches, audit trails, and compliance review.
• Judgment about which research work to automate and which must remain with a human.

Additional qualifications

• Rust, C++, or Go production experience is a plus; Rust is especially valuable.
• Trading experience, ideally market making, HFT, or crypto, is a plus; market microstructure knowledge is beneficial.
• MCP server design, on-prem open-weight model serving, Delta Lake, Dagster, MLflow, GitOps, or Terraform experience is beneficial.

Benefits

• International environment where English is the main language.
• Team events and offsites, a Learning and Development budget, and health insurance for employees and legal dependents.

Impact measures and ways of working

Recruitment process

• The process includes an initial talent-acquisition screening, technical screen, quant interview, and roundup interviews.
• Candidates may request reasonable adjustments to participate in the interview or assessment process.