Hands‑on Senior Data Scientist converting ambiguous scientific/business opportunities into measurable AI product hypotheses, experiments, and working solutions.
Partner with AI, Data, App/Cloud, Frontend engineers, product owners, and domain experts to build/evaluate AI systems across R&D, Commercialization, Manufacturing, and Enabling Functions.
Key Responsibilities
Frame questions into AI hypotheses, success metrics, evaluation plans, and rapid experiments; contribute to agile AI Accelerator cycles.
Build prototypes using Python, SQL, notebooks, APIs, and AWS‑aligned data services.
Support sandboxed/non‑production data problem solving (branch/transform/test/audit code+data experiments).
Evaluate/curate analytical context (instructions, memory, tools, warehouse context, curated source meaning) and build analytical features (embeddings/classifiers/ranking/recommendations/simulation/optimization).
Partner with Data Engineers on datasets, retrieval corpora, metadata, and feature pipelines (S3, Athena, PostgreSQL/RDS, vector DBs, knowledge graphs).
Design evaluations for LLM/RAG/agentic workflows; create rubrics/golden datasets, validate structured outputs, taxonomy hallucination risk, SME review loops.
Use LangGraph/LangSmith/PydanticAI (or similar) to test agent behavior and reliability; assess context vs raw retrieval.
Define KPIs/measurement plans; use demos/sprint reviews to assess MVP progress; apply statistical/experimental/causal or quasi‑experimental methods.
Create analyses/visualizations/narratives explaining behavior, limitations, and opportunities; partner on responsible AI/privacy/governance.
Qualifications & Required/Preferred Skills
BS+ in Data Science/Statistics/CS/Engineering/Bioinformatics/Computational Biology/Applied Math or related.
5+ years in data science/ML/applied AI/analytics.
Proficient in Python, SQL, R; pandas/NumPy/scikit‑learn/PyTorch/TensorFlow/statsmodels.
Familiarity with AWS services (S3, Athena, RDS/PostgreSQL, OpenSearch, SageMaker, Bedrock) and vector DBs/knowledge graphs/embeddings.
Experience with evaluation rubrics, hallucination risk, causal inference, simulation/optimization/recommendations; Streamlit preferred for prototyping.
Communicate findings to technical/non‑technical audiences; use coding agents (Claude Code/Codex/Gemini CLI/Copilot).
Benefits
Health coverage (medical/pharmacy/dental/vision), wellbeing support (accounts/EAP), financial protection (401(k), disability, life/accident/supplemental insurance, travel protection, legal support, identity theft).
Paid time off; US exempt employees: flexible time off (unlimited) + 11 paid national holidays; other listed groups: 160 hours annual vacation + holidays.