Senior Data Engineer — Translational Data Products

Bristol Myers Squibb • Tampa, FL

1d • $107,870 - $143,778 • Hybrid

About The Position

Our team builds and maintains data products that make R&D data usable and AI-ready: biomarker, biospecimen, clinical trial, omics, among others, assembled into governed, curated, linked, documented, AI-ready products that translational scientists and analysts use for insight generation and decision making. You will own data products end to end — from documenting where the data lives, to unlocking the source system with its data owner, through entity mapping, modeling and validation, to making sure the product answers high-value scientific questions and gets used or retired.
You will work alongside data scientists, machine learning engineers, and translational medicine stakeholders, and you will use Claude Code as a normal part of how you deliver.

Requirements

• Requires a Bachelor's degree in a relevant discipline along with 4+ yrs of relevant work experience.
• Demonstrated experience building production data pipelines (ETL/ELT) at scale.
• Very strong SQL — complex joins, window functions, performance tuning, and the judgment to know when a query result is wrong.
• Python for data engineering and data science.
• AWS experience — S3, IAM, Glue, Lambda, ECS and Fargate.
• Containerized service deployment — package a Python service in Docker and ship it to ECS/Fargate through infrastructure as code (AWS CDK / CloudFormation) and a GitHub Actions pipeline, with health checks, secrets from parameter store, and SSO / OIDC authentication. Our MCP servers, which expose data products to AI agents, are deployed this way and you will maintain your own.
• CloudOps / DevOps practice: monitoring, alerting, incident response.
• GitHub fluency: branching, pull requests, code review, and GitHub Actions.
• Test-driven and validation-driven — unit tests that run in the pipeline, plus regression and reconciliation checks that run against the platform (a Databricks job or SQL check) so a pull request fails when a number changes. Point to a validation suite you built and say what it caught.
• Command-line fluency — git, shell, the AWS and Databricks CLIs, and daily hands-on use of Claude Code to drive a task end to end. Claude Code is the approved agentic coding tool here; other agents are not permitted at this time.
• Life sciences data — hands-on experience with clinical, molecular and assay data: patient-level clinical data and lab results, molecular data (genomics, transcriptomics, proteomics), and assay output (biomarker panels, flow cytometry). You know how these data types link to a patient and where they are commonly misinterpreted. Translational or biomarker work is welcome but not expected.
• Ability to engage with the science. You work directly with translational scientists and data owners without a translator: following the biology and study design well enough to ask the next question, and continuing to ask until you know what decision the data has to support. You say so when the question as asked will not answer the need, and you trace data upstream to its original source of truth — how it entered the organization, which system captured it first, what was changed along the way, and who owns it.

Nice To Haves

• Translational medicine, biomarker sciences, or biospecimen domain experience specifically.
• Fluency in clinical trial data standards — SDTM/ADaM, EDC — and comfort with exploratory / non-CDISC research data where no standard applies.
• Databricks — Unity Catalog, Delta Lake, Lakeflow/DLT pipelines, Databricks SQL.
• Depth in a specific omics modality (genomics, transcriptomics, proteomics, flow cytometry), or experience with digital pathology / imaging data.
• Experience extending Claude Code, not only using it — hooks, guardrails, skills, subagents, or automated review steps you built for a team to reuse.
• Experience partnering with machine learning engineers on feature pipelines or RAG/GenAI data preparation.
• Data governance and metadata management experience (lineage, data contracts, cataloging, access models).

Responsibilities

• Build, operate and own data products. Design, build, test, and maintain ETL/data normalization pipelines on Databricks and AWS that create governed, well-documented datasets in Unity Catalog (both tables and views).
• Unlock data access. Work directly with data owners and governance partners to bring new sources under governance and make them AI-ready — cataloged, described, permissioned, quality-checked, and discoverable. This may include csv, tsv, json, xlsx, powerpoint, structured documents, etc.
• Pursue business impact. Engage stakeholders in translational medicine, biomarker sciences, predictive medicine, reverse translation and clinical development to understand and translate the high value questions they are trying to answer into robust data product designs and data maps linking sources to targets (STTM).
• Translate between business and data. Turn a scientific or business question into a data model and a mapping from the question to the SQL query; carefully describe data constraints / context and considerations into caveats that a stakeholder can understand.
• Validate everything. Bring a test-driven, validation-driven mindset: expectations and data quality checks in the pipeline, reconciliation against source, unit and regression tests that run automatically in GitHub Actions on every pull request, and documented evidence that a number is correct before anyone depends on it.
• Engineer for robustness. Create the hooks, guardrails, CDK / CloudFormation infrastructure as code, and automated checks that keep our work reliable and aligned with BMS enterprise IT standards — SDLC, security, SSO / identity and access management, mandatory resource tagging and service registration (CRID, ISR assessment), code review, secrets handling, change management. Run the required cyber checks — dependency, container and infrastructure-as-code scanning (Wiz, Dependabot) on every pull request — and respond quickly when a vulnerability is reported, remediating critical findings within BMS Cybersecurity timelines.
• Work with ML engineers. Prepare feature-ready and model-ready datasets, and support ML and GenAI workloads built on top of our products; support deployment and accessibility of data products on NVIDIA clusters.
• Work as part of a team. GitHub-based development, pull requests, code review, automated build/test/deploy pipelines (continuous integration and delivery, CI/CD), infrastructure as code.
• Continuous improvement. Work with stakeholders to ensure the data products continue to meet their expectations as their work evolves and needs change.

Benefits

• Health Coverage: Medical, pharmacy, dental, and vision care.
• Wellbeing Support: Programs such as BMS Well-Being Account, BMS Living Life Better, and Employee Assistance Programs (EAP).
• Financial Well-being and Protection: 401(k) plan, short- and long-term disability, life insurance, accident insurance, supplemental health insurance, business travel protection, personal liability protection, identity theft benefit, legal support, and survivor support.
• Paid Time Off
• US Exempt Employees: flexible time off (unlimited, with manager approval, 11 paid national holidays (not applicable to employees in Phoenix, AZ, Puerto Rico or Rayzebio employees)
• Phoenix, AZ, Puerto Rico and Rayzebio Exempt, Non-Exempt, Hourly Employees: 160 hours annual paid vacation for new hires with manager approval, 11 national holidays, and 3 optional holidays
• Based on eligibility, additional time off for employees may include unlimited paid sick time, up to 2 paid volunteer days per year, summer hours flexibility, leaves of absence for medical, personal, parental, caregiver, bereavement, and military needs and an annual Global Shutdown between Christmas and New Years Day.
• All global employees full and part-time who are actively employed at and paid directly by BMS at the end of the calendar year are eligible to take advantage of the Global Shutdown.
• The summer hours program is for United States (U.S.) office-based employees due to the unique nature of their work. Summer hours are generally not available for field sales and manufacturing operations and may also be limited for the capability centers. Employees in remote-by-design or lab-based roles may be eligible for summer hours, depending on the nature of their work, and should discuss eligibility with their manager. Employees covered under a collective bargaining agreement should consult that document to determine if they are eligible. Contractors, leased workers and other service providers are not eligible to participate in the program.

What This Job Offers

• Job Type: Full-time
• Career Level: Senior