Data Scientist - Full-time - Worldwide

Data Scientist

The Role

We’re looking for a Data Scientist who can turn messy, real-world insurance data into reliable signals that power our AI assistants. This is not a “research only” role — you’ll be building evaluation frameworks, data pipelines, and applied models that go into production with some of the largest insurers in the world.
You’ll work closely with engineers and product leads to push the boundaries of what AI can do in insurance, while ensuring our solutions are trustworthy, compliant, and performant.

What You’ll Do

• Data Ingestion & Modeling: Design pipelines to process structured and unstructured insurance data (e.g., census files, policy docs, ACORD forms, filings).
• Agent Training & Evaluation: Build evaluation frameworks for Stitch’s micro-agents — deterministic, repeatable, and scenario-based.
• Entity & Feature Extraction: Identify and normalize benefits, exclusions, and policy rules for downstream AI workflows.
• Model Experimentation: Leverage LLMs, embeddings, and RAG techniques to improve retrieval and contextual reasoning.
• Metrics & Validation: Define and track key performance metrics (precision, recall, coverage) across agents.
• Collaboration: Partner with engineering and product to embed data science into Stitch’s platform, with compliance and security as first-class priorities.

What We’re Looking For

Core Skills

• Strong foundation in applied machine learning, NLP, or information retrieval.
• Real-world production experience building and deploying AI/LLM-powered systems (not just prototypes or research).
• Proficiency with Python, Pandas, and SQL; experienced in data wrangling at scale.
• Experience designing evaluation frameworks or structured output pipelines.
• Hands-on work with embeddings, vector databases, and retrieval-augmented generation.
• Ability to combine first-principles thinking with a bias for fast, pragmatic delivery.
• Ability to explain how the attention mechanism in transformers works, and how that understanding influences prompt design, context structuring, and debugging in real-world LLM applications

Nice to Have

• Background in insurance or financial services data.
• Familiarity with compliance-heavy environments (SOC 2, HIPAA, GDPR).
• Cloud experience (AWS or Azure) and exposure to MLOps.
• Knowledge of LangChain, LangGraph, or orchestration frameworks.

Who You Are

• You thrive in a fast-moving startup environment and are comfortable with ambiguity.
• You care about building trustworthy, production-ready AI systems — not just demos.
• You like working across disciplines and see data science as a team sport.
• You’re curious, rigorous, and confident enough to challenge assumptions.

Why Join Stitch Studio

• Work on mission-critical AI for an industry ready for transformation.
• Join a small, senior team where your impact is immediate and visible.
• Help define the evaluation standards and data pipelines that make AI practical.
• Competitive startup compensation (salary + equity) with flexibility on location.