Senior Software Engineer — RL Environments & Infrastructure

Lumos; San Francisco (CA); On-site; USD 162,000 - 198,000; Full time

Job summary

Lumos is a health and life-science AI evaluation company based in San Francisco. We are expanding our backend engineering team to build and operate scalable RL environments on Google Cloud using Python, turning research ideas into reliable production systems.

The role focuses on designing, implementing, and maintaining the infrastructure that runs rewards, verifiers, and evaluation logic, collaborating closely with research, data, and product teams.

Qualifications

• Senior backend engineer with 5+ years of experience
• Strong Python expertise
• Hands-on with Google Cloud Platform
• Experience integrating third-party APIs into production systems
• Familiar with modern engineering practices: CI/CD, testing, observability

Responsibilities

• Build and operate RL environments and harnesses at scale
• Implement verifier and grader plumbing based on research designs
• Maintain infrastructure on Google Cloud: compute, storage, networking, deployment
• Integrate model providers (OpenAI, Anthropic, Google, others) with attention to cost and latency
• Write clean, well-tested code and set standards through code reviews
• Collaborate with research, data, and product teams to ship running systems

Skills

Python

Backend engineering

Google Cloud Platform

CI/CD

Testing

Observability

Version control

API integration

Go

Job description

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 scaling our Paris team to support new product launches and foundational lab partnerships - startup velocity, frontier-model stakes, in one of the most beautiful cities in the world.

The Role

You'll build and operate the systems that run our RL environments at scale - the harnesses, tool sandboxes, verifier plumbing, and rollout infrastructure that turn a reward-design idea into a working environment our partners can train against. This is the engineering counterpart to our research team: they design the tasks, rewards, and eval logic; you make them real, reliable, and fast.

This is a deep backend role on Google Cloud, with Python as the primary language.

Our Research Scientists design reward, verifier, and evaluation logic and own the research questions. You build, scale, and operate the environments and infrastructure that execute it.

What You Will Do

• Build and operate RL environments and the harnesses that run them - task execution, tool sandboxes, state management, and rollout orchestration.
• Implement verifier and grader plumbing designed by the research team - turning scoring logic into reliable, reproducible, fast code.
• Build and maintain the infrastructure these environments run on, on Google Cloud (compute, storage, networking, deployment).
• Integrate and orchestrate model providers (OpenAI, Anthropic, Google, and others), with attention to evaluation, cost, and latency.
• Write clean, well-tested, maintainable code, and set engineering standards through review.
• Collaborate closely with research, data, and product teams to turn environment designs into shipped, running systems.

What You Bring

• Senior backend engineering experience (typically 5+ years), with Python as a primary language.
• A proven track record owning systems end to end - from design through production operation.
• Hands-on experience with Google Cloud Platform (compute, storage, networking, deployment).
• Experience integrating third-party APIs and AI / LLM vendor services into production systems.
• Comfort with modern engineering practices: version control, CI/CD, testing, observability.

Nice to Have

• Experience building or working with reinforcement-learning systems (directly relevant to our RL / RLAIF work).
• Hands-on experience with agentic AI and AI-assisted development - using coding agents as part of real production work.
• Experience with Go.
• Experience scraping and ingesting data from diverse external sources.
• Interest in or exposure to health / life-science domains.

Compensation

• From $180,000 base salary
• Annual performance bonus
• 20,000-50,000 stock options

Work Authorization

Candidates must be authorized to work in the US without sponsorship. Lumos does not sponsor work authorization for this role.

Lumos is an equal opportunity employer. We are committed to building a diverse and inclusive team and do not discriminate on the basis of race, color, religion, gender, gender identity or expression, sexual orientation, national origin, genetics, disability, age, or any other characteristic protected by applicable law. All qualified applicants will receive consideration for employment.