Founding AI Engineer
Founding AI Engineer
A Boston-based stealth startup is building an AI‑powered operating system for cancer screening, targeting early detection in healthy populations and recurrence monitoring for survivors. The team is made up of technical leaders recognized for sequence‑modeling breakthroughs and large‑scale decision systems deployed across high‑stakes domains. This group has recently secured a significant seed round and is building it's founding technical team.
The company is building genome scale foundation models combining large‑scale genomics and longitudinal medical data with simulation‑driven decisioning to drive better patient outcomes.
The Role:
This hire will help set the foundational direction for the AI engineering effort and contribute to building real systems quickly. The role comes with meaningful ownership, fast iteration cycles, and a high bar for technical quality.
What you'll work on
- Core model training and experimentation workflows: contributing to the design and evolution of large‑scale training pipelines for foundation models across genomics and medical data.
- Scaling and efficiency: improving training performance, system efficiency, and cost awareness as models and datasets grow, with opportunities to go deep on optimization over time.
- Simulation and decision-making systems: helping integrate model outputs into downstream simulation or optimization loops that support personalized screening strategies.
- Evaluation and robustness: developing benchmarks, tests, and monitoring to ensure models behave predictably and improvements are measurable.
- Early architectural decisions: participating in foundational design choices that influence modeling velocity, system reliability, and long‑term scalability.
Preferred qualifications
- 3-5 years of hands‑on experience building ML systems (industry or research‑heavy).
- Strong JAX + Python, comfort with distributed training, XLA compilation, and performance instrumentation.
- Evidence of performance engineering: you've squeezed real speed/scale out of real hardware.
- Nice‑to‑have: genomics/clinical data exposure
- Prefers in‑person, small‑team intensity; high ambition and ownership.