Principal Data Scientist - RPL
Summary
Principal Data Scientist builds and scales causal-inference models for promotion optimization, uplift modeling, and budget-constrained allocation in a large Southeast-Asian marketplace.
What You Will Do
- Own the end-to-end modelling and estimation behind promotion optimisation — elasticity, heterogeneous treatment effects, budget-constrained allocation, and increamentality — from problem framing to production.
- Advance our causal inference stack: experiment and quasi-experiment design (geo/switchback tests, holdouts, diff-in-diff, synthetic control), debiasing observational data, and variance reduction — raising the bar on experimentation rigor across the team.
- Design and productionize heterogeneous treatment effect (uplift) models at scale (tens of millions of users), with honest offline evaluation (uplift/Qini curves, policy-value estimation).
- Formulate and solve budget-constrained allocation under fairness and dynamic business constraints — from LP/MILP to greedy or Lagrangian methods where they scale better.
- Mentor and technically guide a team of data scientists; set standards for model evaluation, documentation, and scientific review.
- Partner with business to turn model outputs into budget decisions, and communicate tradeoffs (subsidy efficiency vs. growth) clearly.
What You Will Need
- 8+ years in data science or ML, with 3+ years focused on causal inference or uplift modelling in production settings.
- Deep expertise in heterogeneous treatment effect estimation, with hands-on production experience in several of: meta-learners (S/T/X/R), causal forests, DR-learner, or deep uplift architectures.
- Hands-on experience optimising promotions, pricing, or marketing incentives with evolving constraints and measurable business outcomes.
- Strong grounding in experimentation and observational causal methods — propensity weighting, instrumental variables, synthetic control, difference-in-differences.
- Experience with constrained optimisation (MILP, Lagrangian methods, etc.) applied to resource allocation.
- Proficiency in Python and SQL; shipping models to production with engineering partners.
- Track record of technical leadership at principal/staff level: setting technical direction for a team, reviewing high-stakes analyses, and influencing roadmaps and partner teams without direct authority.
- Experience in marketplaces, ride-hailing, food delivery, e-commerce, or fintech.