Scientific Project Lead, Autonomous Lab

About Lumos
Lumos is a health and life-science AI evaluation company based in San Francisco. Our mission is to make health and life-science AI genuinely accurate and safe. We build the benchmarks, RL environments, and adversarial testing frameworks that frontier labs rely on to know whether their models can be trusted with science.

We’re opening a position in our Paris team to support new product launches in autonomous experimentation. You get startup speed and frontier-model stakes, working in the middle of the AI ecosystem.

The Role
You’ll be the lab automation expert at the center of how Lumos builds RL tasks for autonomous wet labs. This is a hands-on, full-time role based in Paris. The core of the work is:

• Designing and delivering RL tasks and environments where AI agents plan, run, and interpret wet-lab experiments on real or simulated automation hardware.

• Turning real liquid-handling and autonomous-lab workflows into gold-standard tasks, reference solutions, and automated verifiers that become training and evaluation data for frontier models.

• Owning delivery end to end: scoping with customers, building with our research and engineering team, and holding the scientific and technical quality bar.

You’ll define what “correct,” “safe,” and “executable” mean for an AI agent driving a lab. You’ll find where models fail, and you’ll be the credible lab automation voice for customers and our expert network.

What You Will Do

• Build RL tasks and environments around autonomous lab workflows. This covers protocol design and translation to robot-executable code, liquid-handling method development (pipetting strategies, liquid classes, deck layouts, labware definitions), plate-based assays, serial dilutions and normalizations, PCR and NGS library prep, cell culture and screening workflows, and closed-loop design-make-test-analyze cycles.

• Write and review reference solutions and verifiers that check agent outputs automatically. That includes protocol correctness, volume and labware constraints, deck-state validity, contamination and error handling, and whether a protocol would actually run on hardware.

• Integrate with lab automation APIs and simulators (Opentrons Python API, Hamilton VENUS / PyHamilton / PyLabRobot, scheduling and orchestration layers) to build environments agents can act in.

• Stress-test and red-team models on lab tasks. Find where they hallucinate protocols, violate physical constraints, or produce unsafe or non-executable instructions, and turn those failures into lasting evaluations.

• Lead delivery on customer projects. Scope tasks, manage timelines, review the work of a network of scientific and automation experts, and own inter-rater reliability and rubric quality.

• Be the credible lab automation voice with customers (frontier labs and AI-for-science companies) and with the expert network you help recruit and calibrate.

What You Bring

• PhD or MSc in biology, chemistry, bioengineering, or a related field, or equivalent industry experience.

• Hands-on experience with liquid-handling robots, including significant work on Hamilton (STAR/STARlet/Vantage, VENUS) and/or Opentrons (OT-2/Flex, Python API).

• A track record of setting up and running automated or autonomous lab workflows end to end, from protocol design to integration, validation, and troubleshooting on real hardware.

• Strong coding skills in Python. You’re comfortable writing production-quality code, working with REST APIs and SDKs, and integrating instruments, schedulers, and data pipelines.

• A solid grasp of wet-lab fundamentals and experimental design. You know the difference between a protocol that looks right and one that actually runs and gives a reproducible result.

• Rigor about correctness, reproducibility, and edge cases.

• Experience leading projects or teams and delivering against deadlines with external stakeholders.

• Excellent written and spoken communication in English. French is a plus.

Nice to Have

• Experience with lab orchestration and scheduling software or open-source frameworks (PyLabRobot, SiLA 2).

• Experience with other platforms (Tecan, Beckman Biomek, Chemspeed) or integrated instruments (plate readers, incubators, robotic arms).

• Experience with self-driving labs, Bayesian optimization, or active-learning loops for experiment selection.

• Prior experience annotating, evaluating, or red-teaming AI or LLM systems, or using LLMs to generate lab protocols.

• Exposure to RL, verifier-based scoring, or evaluation frameworks.

• Publications or open-source contributions in lab automation or AI for science.

Work Authorization
Candidates must be authorized to work in France / the EU.

Start date: Flexible, ASAP preferred

Equal Opportunity
Lumos is an equal opportunity employer. We are committed to building a diverse and inclusive team and do not discriminate on the basis of origin, sex, gender identity, sexual orientation, family situation, pregnancy, age, disability, health, genetic characteristics, religion, political opinions, union activity, or any other characteristic protected by applicable law. All qualified applicants will receive consideration for employment.