Applied Scientist, AI Product Methods
HybridPalo Alto, California, United States
Job Summary
Define methodologies for AI capabilities such as reasoning, planning, retrieval, and grounding by turning product needs and system limitations into clear methodological questions. Evaluate relevant research and industry techniques to determine what should be adopted, adapted, or rejected, then design approaches with explicit assumptions, workflows, decision rules, and evaluation criteria. Build Python prototypes and reference implementations on Azure, create evaluations measuring effectiveness and failure modes, and compare alternative approaches to make evidence-based recommendations. Convert validated methods into specifications, reference code, and acceptance criteria before partnering with Product, Engineering, and domain experts through integration and refinement.
Required Qualifications
- You have taken an AI idea from concept to something real: a prototype, internal tool, or product feature.
- You design methods with clear assumptions, failure modes, and evaluation criteria.
- Practical Python and Cloud infrastructure experience and the ability to build working prototypes others can run and extend.
- Experience designing or owning experiments, evaluations, or benchmarks that informed a ship/change/kill decision.
- The ability to understand research findings and determine how they should – or should not – be applied in a product.
- A solid understanding of machine learning and modern AI systems (for example language models, retrieval, agents, verification, or feedback loops).
- Clear technical writing: you can turn a validated approach into a specification, reference implementation, and acceptance criteria.
- Comfort working across Product, Engineering, and domain teams in a fast-moving environment.
- Experience with language models, retrieval systems and memory, agents, or production AI evaluation.
- A thesis-based master's degree or PhD in computer science, artificial intelligence, machine learning, control theory, robotics, autonomous systems, or a related field.
Desired Qualifications
- A peer-reviewed publication record (helpful signal).
Hiring someone like this?
Get your role in front of qualified candidates on Sorce.