Software Engineer ML Infrastructure
Independent nonprofit research institute building AI-and-biology tools and conducting biomedical research on complex diseases.
About Arc Institute
Arc Institute is a Palo Alto-based independent nonprofit founded in 2021. It supports long-term biomedical research and develops experimental, computational, and machine-learning technologies, including virtual-cell and genomic foundation models, in partnership with Stanford University, UC Berkeley, and UCSF.
Skills
Candidate Availability
Required and preferred rules are kept separate and reflect the wording in the original posting.
About the Role
You will own ML infrastructure for data, model evaluation, training, and validation. You will build evaluation tools and training-data pipelines, connect dataset versions to model performance, support lab-in-the-loop workflows, validate agent-written software, and improve distributed training reliability, throughput, and GPU utilization.
Requirements
- Bachelor’s degree with 8 years of relevant experience, or a Master’s degree with 6 years of relevant experience.
- Ownership of a production data or ML platform, including pipelines, storage, versioning, and technical direction.
- AI tools and agents
- Evaluation, quality-control gates, or tests for verifying AI-generated output.
- Technical-domain learning and design decision-making.
- Stakeholder requirements gathering.
- Code review, pairing, and technical documentation.
- Python or PyTorch performance debugging using profilers and measurements.
- GPU utilization, memory, and data-loading optimization.
- Work onsite in the Palo Alto office at least 3 days per week.
Responsibilities
- Own how virtual cell model performance is measured, beginning with the Model Performance Dashboard backend.
- Design and build self-service evaluation and model-comparison tools for researchers.
- Redesign and operate the pipeline that turns single-cell data into model training sets.
- Use AI agents to find, process, and curate datasets, supported by automated quality checks.
- Own lab-in-the-loop infrastructure that informs laboratory experiments from model-performance results.
- Build or procure validation frameworks for agent-written code and applications.
- Own engineering for reliable, scalable ML training and evaluation.
- Support multi-node training, diagnose distributed-training failures, and improve throughput and GPU utilization.
Benefits
- Annual discretionary bonus eligibility.
- Participation in the Compound Cash Grant Program, with multi-year cash grants paid in semi-annual installments.