Founding AI Evidence Engineer

Department: Engineering

Location: San Francisco

Employment Type: FullTime

About the Role

This is a founding engineer role at a small, high-impact startup building the evidence infrastructure layer for safety-critical AI — the system that proves a model works, and keeps proving it across its entire lifecycle. The company is focused on diagnostic medical imaging, where assembling rigorous evidence about AI behavior is a genuine and growing pain point for vendors, regulators, and hospitals alike.

As regulations tighten across healthcare, financial services, and autonomous systems, regulated AI vendors face mounting costs assembling evidence for regulators, buyers, and payers — and repeating that process with every model update. This role sits at the center of solving that problem: making evidence generation continuous rather than reactive.

This is one of very few roles where deep technical skill combined with pragmatism translates directly into outsized real-world impact.

What You'll Do

• Work directly with medical imaging companies preparing FDA 510(k) or De Novo submissions, and with hospitals evaluating those models.

• Translate vague customer questions into specific, structured investigations — run end-to-end in one- to two-week cycles.

• Surface failure modes, edge cases, and distribution drift in real-world AI systems.

• Produce evidence artifacts that customers use for development decisions, regulatory go/no-go calls, and hospital governance conversations.

• Codify recurring investigation workflows into reusable methodologies and pipelines that become core product.

• Help shape how the team runs engagements as the company grows — cadences, standards, and hiring.

What We're Looking For

Required:

• Strong empirical ML skills — you've designed investigations, analyzed failure cases, and reasoned rigorously about distribution shift and uncertainty in real systems.

• Demonstrated ability to build reusable infrastructure from one-off work: pipelines, internal tooling, and evaluation frameworks.

• Python fluency across the full stack — from raw data wrangling to defensible, well-communicated conclusions.

Nice to have:

• Experience with medical imaging data (DICOM, radiology modalities, etc.).

• Familiarity with FDA regulatory pathways (510(k), De Novo) or other safety-critical compliance frameworks.

• Background in model evaluation, AI safety, or algorithmic auditing.

• Comfort working in a very early-stage, customer-facing technical role.

Compensation & Benefits

• Salary: $150,000 – $230,000 USD annually

• Visa sponsorship is available

• Founding team role — meaningful equity expected

Location

This is an on-site role based in San Francisco, California, USA. Candidates should be prepared to work in-person.