Translational AI Engineer
Pfizer
September 02, 2026
Remote friendly (Cambridge, MA)
United States
Clinical Research and Development
Role Responsibilities:
- Design and build fit-for-purpose AI/ML systems for recurring scientific and clinical workflows, including hybrid RAG.
- Harden AI-assisted prototypes by identifying technical and scientific failure modes and rebuilding what needs engineering before scaling.
- Stand up lean data platform and ETL underneath these systems (in partnership with Digital).
- Own internal AI tool promotion from alpha to beta by defining “production-ready” and building evaluation harnesses.
- Establish and document SOPs for rigorous build practices from day one.
- Rigorously evaluate commercial AI/ML and GenAI tools/vendors against a build-vs-buy bar (capability, cost, security, scientific fit, workflow readiness).
- Run evaluation loops using workflow-specific scientific/clinical benchmarks with input from computational biologists, immunologists, biologists, and clinicians.
- Distill learnings into hardened primitives, reference architectures, and benchmark harnesses.
- Apply scientific-quality rigor: fit-for-purpose evaluation, grounded outputs, documentation, guardrails, disclosure of model limitations, and human oversight.
- Stay current on translational AI methods/infrastructure (generative AI, agentic systems, predictive modeling, foundation models, retrieval/fine-tuning, evaluation).
- Contribute to a rigor-focused culture and build external presence via conferences and publications.
Basic Qualifications:
- PhD (1+ years) OR Master’s (5+ years) OR Bachelor’s (6+ years) shipping AI/ML systems for scientific applications.
- Hands-on AI/ML depth in at least one modality (genAI, agentic workflows, predictive models, foundation models, hybrid RAG, fine-tuned architectures).
- Ability to evaluate/debug/harden AI-assisted codebases.
- Strong Python with tools/frameworks including PyTorch, HuggingFace, LangChain, or LlamaIndex.
- In-depth database/ETL experience (Postgres or equivalent).
- Cloud/deployment fluency (AWS/GCP/Azure), containerization, CI/CD.
- Translational research/clinical workflow literacy.
- Demonstrated build-vs-buy judgment.
- Experience building evaluation harnesses and error analysis into systems.
- Ability to deliver in shifting-requirement environments.
Preferred Qualifications:
- Took an internal tool from “mostly works” to hardened, evaluated system.
- Shipped fit-for-purpose systems over messy scientific/clinical data.
- Conducted build-vs-buy evaluations that held up over time.
- Full-stack capability beyond model layer (TypeScript/React or equivalent).
- Experience with immunology/inflammation and/or multi-omics data.
- Familiarity with biological foundation models (e.g., scGPT, Geneformer, ESM, AlphaFold).
- Built production enterprise software with security/compliance/auditability/uptime.
- Strong open-source/publication record.
Application Instructions / Job Details:
- Last date to apply: September 16, 2026.
- Hybrid role: live within commuting distance; work on-site ~2.5 days/week.
- Design and build fit-for-purpose AI/ML systems for recurring scientific and clinical workflows, including hybrid RAG.
- Harden AI-assisted prototypes by identifying technical and scientific failure modes and rebuilding what needs engineering before scaling.
- Stand up lean data platform and ETL underneath these systems (in partnership with Digital).
- Own internal AI tool promotion from alpha to beta by defining “production-ready” and building evaluation harnesses.
- Establish and document SOPs for rigorous build practices from day one.
- Rigorously evaluate commercial AI/ML and GenAI tools/vendors against a build-vs-buy bar (capability, cost, security, scientific fit, workflow readiness).
- Run evaluation loops using workflow-specific scientific/clinical benchmarks with input from computational biologists, immunologists, biologists, and clinicians.
- Distill learnings into hardened primitives, reference architectures, and benchmark harnesses.
- Apply scientific-quality rigor: fit-for-purpose evaluation, grounded outputs, documentation, guardrails, disclosure of model limitations, and human oversight.
- Stay current on translational AI methods/infrastructure (generative AI, agentic systems, predictive modeling, foundation models, retrieval/fine-tuning, evaluation).
- Contribute to a rigor-focused culture and build external presence via conferences and publications.
Basic Qualifications:
- PhD (1+ years) OR Master’s (5+ years) OR Bachelor’s (6+ years) shipping AI/ML systems for scientific applications.
- Hands-on AI/ML depth in at least one modality (genAI, agentic workflows, predictive models, foundation models, hybrid RAG, fine-tuned architectures).
- Ability to evaluate/debug/harden AI-assisted codebases.
- Strong Python with tools/frameworks including PyTorch, HuggingFace, LangChain, or LlamaIndex.
- In-depth database/ETL experience (Postgres or equivalent).
- Cloud/deployment fluency (AWS/GCP/Azure), containerization, CI/CD.
- Translational research/clinical workflow literacy.
- Demonstrated build-vs-buy judgment.
- Experience building evaluation harnesses and error analysis into systems.
- Ability to deliver in shifting-requirement environments.
Preferred Qualifications:
- Took an internal tool from “mostly works” to hardened, evaluated system.
- Shipped fit-for-purpose systems over messy scientific/clinical data.
- Conducted build-vs-buy evaluations that held up over time.
- Full-stack capability beyond model layer (TypeScript/React or equivalent).
- Experience with immunology/inflammation and/or multi-omics data.
- Familiarity with biological foundation models (e.g., scGPT, Geneformer, ESM, AlphaFold).
- Built production enterprise software with security/compliance/auditability/uptime.
- Strong open-source/publication record.
Application Instructions / Job Details:
- Last date to apply: September 16, 2026.
- Hybrid role: live within commuting distance; work on-site ~2.5 days/week.