Data Science Manager

Start free →

Or apply yourself on the official page →

About the role

Work Location Assignment:  Mexico City, must be able to work from assigned Pfizer office 2-3 days per week, or as needed by the business. ROLE SUMMARY Applied Intelligence is a high-velocity team purpose-built to do one thing exceptionally well: take the hardest, most ambiguous AI/ML problems in the enterprise and rapidly determine whether they are solvable, how they should be solved, and what it will take to make them real at scale. Our work de-risks high-impact AI/ML investments through fast, disciplined experimentation, and directly shapes what gets scaled, what gets stopped, and where the organization invests next. As Manager you will build and lead a small team of exceptional applied AI engineers and data scientists while staying deeply hands-on yourself, contributing directly to architecture, code, and experimental design. This is not a sandbox: every proof of concept your team builds is developed with the engineering hygiene of production code, because the best prototypes become the foundation of enterprise systems. The role rests on three pillars, and we expect strength in all three: business judgment to choose and frame the right problems, technical depth to build and validate AI/ML solutions, and engineering discipline to make the work reproducible, reviewable, and ready to scale. Day to day, you will work in 2 to 6 week prototype cycles, translating ambiguous commercial problems into testable hypotheses, building models and data pipelines, and delivering evaluation evidence that supports clear go/no-go decisions.    ROLE RESPONSIBILITIES Business Skills and Problem Framing • Partner with commercial, product, and functional leaders to identify where AI/ML can create real value, and to say clearly when it cannot. • Translate ambiguous business problems into scoped, testable hypotheses with defined success criteria before any code is written. • Deliver clear, outcome-oriented recommendations on what to scale, what to stop, and where to invest next, with the evidence and the limitations stated plainly. • Communicate results to technical and non-technical audiences alike, including senior stakeholders, and defend methodological choices under scrutiny. • Manage a portfolio of concurrent prototypes, balancing speed against rigor and making explicit trade-off calls on scope and effort. Building AI/ML Models • Contribute hands-on to model development: feature engineering, model selection, training, tuning, and evaluation across classical ML, deep learning, and LLM-based approaches. • Own experimental design with statistical and methodological rigor, including validation strategy, baselines, evaluation metrics, error analysis, and avoidance of leakage and other common pitfalls. • Build data pipelines and integrations that make prototypes real, working with enterprise, third-party, and unstructured data sources. • Apply GenAI and LLM techniques where they fit the problem, including retrieval-augmented generation, agentic workflows, prompt engineering, and systematic evaluation of model outputs. Engineering Discipline • Enforce GitHub-based workflows as a baseline: branching strategy, pull requests, code review, and traceable history on every project. • Build and maintain CI/CD pipelines for models and data products, including automated testing, linting, and reproducible builds. • Deliver reproducible, production-ready code: containerization, dependency and environment management, configuration over hardcoding, and clear documentation. • Apply MLOps practices to produce deployment-ready artifacts, including experiment tracking, model versioning, monitoring, and rollback strategies appropriate for regulated environments. • Use AI coding tools (for example Claude Code, Cursor, GitHub Copilot) to accelerate delivery an

View the official posting →