Senior Software Engineer — Data Platform & Infrastructure

Lumos AI Research is seeking a Senior Software Engineer to build the data platform and infrastructure that supports the end-to-end lifecycle of high-quality datasets for frontier AI models, from creation and curation through evaluation and delivery. The role spans product systems and infrastructure, with ownership across workflows, AI-assisted interfaces, orchestration, quality evaluation, and scaling.

Responsibilities

• Design and implement core product capabilities that allow experts to create, annotate, review, validate, and refine complex training and evaluation datasets.
• Build multi-stage pipelines that combine human contributions, AI-generated content, automated evaluation, and expert review.
• Create collaboration tools that help experts work with AI models, including AI-assisted annotation, content generation, feedback, and validation.
• Develop robust dataset systems covering storage, versioning, lineage, transformations, quality control, and delivery.
• Implement workflow mechanics such as assignment, routing, dependencies, review queues, quality gates, and workflow automation across distributed expert teams.
• Integrate model-based judges, programmatic verifiers, consensus mechanisms, and quality metrics to improve data reliability.
• Build fast, intuitive interfaces for complex scientific and healthcare workflows while balancing flexibility with simplicity.
• Develop infrastructure for model inference, structured generation, prompt execution, and automated feedback loops.
• Improve performance, reliability, observability, and developer experience as data volume, workflow complexity, and expert participation grow.
• Drive technical decisions from product design and system architecture through implementation, deployment, and iteration.

Requirements

• 5+ years of software engineering experience, or equivalent evidence of exceptional engineering ability.
• Strong proficiency in TypeScript and Python, including experience building sophisticated, production-grade applications.
• Experience with modern frontend frameworks such as React or Next.js, plus a strong understanding of frontend architecture and application state management.
• Solid backend engineering experience, including APIs, relational databases, asynchronous processing, and distributed services.
• Experience designing complex data models, workflow systems, and scalable application infrastructure.
• Strong knowledge of PostgreSQL, queues, background jobs, caching, and cloud infrastructure.
• Ability to build user experiences for technically complex products.
• Strong product intuition and the ability to convert ambiguous requirements into maintainable solutions.
• Demonstrated capability to independently architect, ship, and iterate on complex systems.

Technologies

• TypeScript
• Python
• React
• Next.js
• PostgreSQL

Compensation and Location

Location: San Francisco, CA (onsite)

Salary: USD 160,000 - 220,000 per yearly

Benefits

• Competitive compensation and meaningful equity: Participate in the long-term value you help create.
• Freedom to experiment: Prototype quickly, challenge assumptions, and pursue better approaches without unnecessary bureaucracy.
• Compensation Range: $160K - $220K

What We’re Looking For

• Experience building data annotation, labeling, evaluation, or human-in-the-loop platforms.
• Experience with AI training data pipelines, dataset curation, synthetic data generation, or model evaluation.
• Familiarity with LLM APIs, structured generation, tool calling, prompt management, and AI-assisted workflows.
• Experience building workflow orchestration systems, task queues, or configurable workflow engines.
• Knowledge of data versioning, provenance, lineage, schema validation, and reproducibility.
• Experience designing collaborative tools involving review, approvals, permissions, and complex state transitions.
• Background in developer tools, internal platforms, scientific software, or other technically demanding products.
• Experience working with healthcare, life sciences, or scientific datasets.
• A strong preference for intellectually curious, independent builders who enjoy difficult problems, move quickly, and take pride in exceptional software.

What We Offer

• Develop systems that convert human expertise and scientific knowledge into high-quality data for training and evaluating advanced AI models.
• Shape the architecture, user experience, and evolution of a core platform from early design through production.
• Work across human-AI collaboration, workflow orchestration, data systems, model integration, and quality infrastructure.
• Collaborate with researchers, scientists, and domain experts to understand model failures and the data required to improve outcomes.
• Create tools used by physicians, scientists, and researchers to produce and validate complex data.
• Work directly with founders and exceptional engineers in a fast-moving, intellectually demanding environment.
• Prototype quickly, challenge assumptions, and pursue better approaches without unnecessary bureaucracy.
• Participate in the long-term value you help create.
• Help define engineering culture, product architecture, and the future of AI data infrastructure.

About Lumos AI Research

Lumos AI Research is building the infrastructure to train and evaluate frontier AI systems in healthcare and life sciences. As AI capabilities increase, progress depends not only on better models, but on the quality of the data, environments, and feedback used for training. The team builds infrastructure that supports clinical reasoning, genomics, drug discovery, and laboratory automation, combining deep domain expertise with software infrastructure to generate training data, evaluate model behavior, and develop reinforcement learning environments for increasingly autonomous AI systems.

Lumos is a small, ambitious team of engineers, scientists, and researchers expanding what AI can accomplish in the real world.