Founding Technical Staff

## About the role As a Founding Member of Technical Staff, you'll help build the reasoning methodology for safety-critical AI validation. You'll investigate how models behave, where validation breaks down, and how claims about model behavior can be supported, challenged, or refined with evidence. ## Key Responsibilities - Design and execute investigations into how models perform and fail. Analyze input data, model outputs, and internal representations to evaluate data quality, generalization limits, distribution shift, subgroup performance, and demographic bias. Surface failure modes and produce evidence that supports, challenges, or refines claims about model performance and safety. - Develop Tessel's evidence methodology. Define how claims, arguments, and evidence should be structured for rigorous AI validation. Pressure-test assumptions, critique weak argument structures, and turn repeated investigations into reusable methods, workflows, and platform primitives. - Contribute to shipping the platform by writing production-quality Python, building agentic workflows for evidence investigation, and prototyping front-end features with AI tooling. This is a founding role. We're looking for someone who can challenge our assumptions, contribute beyond their immediate technical domain, and help build foundational infrastructure for the future of safety-critical AI. ## Skills & Experience ### Required - Degree in CS, math, physics, engineering, or a related quantitative field, or equivalent demonstrated depth. - Strong ML, statistics, and data science fundamentals, with an ability to understand the math behind methods, recognize broken mathematical assumptions, and reason about trade-offs. - Expert Python skills, from raw data analysis to writing platform code others will use. - Strong engineering judgment on code structuring, interface boundaries, and trade-offs related to reusability. - Comfort using AI tooling as a primary mode of working. ### Nice-to-Have - 3–5 years of professional ML experience or a PhD in model evaluation, robustness, OOD detection, interpretability, or a related area. - Experience with AI/ML medical device submissions, FDA review processes, or other regulated environments. - Background in safety case methodology in aviation, automotive, healthcare, or other safety-critical fields. ## About the company ### About Tessel Tessel is building the evidence infrastructure for safety-critical AI: the system that proves a model works and keeps proving it. The biggest challenge in safety-critical sectors isn't building models that beat benchmarks. It's producing and structuring the evidence to continuously demonstrate to regulators, governance boards, patients, and insurers that a model is safe and effective in real-world use. We partner with diagnostic imaging AI companies and academic medical centers to rethink how validation evidence is produced and used today. By bringing the rigor of safety cases to AI, we make assumptions visible, surface failure modes, and turn validation work into a systematic process. The result is a platform for continuous monitoring and improvement, one that iteratively closes the gap between validation and patient outcomes.