Principal Data Scientist / Simulation Engineer

MSc or higher in Statistics, Applied Mathematics, or Computer Science · 5+ years in

production modeling

Responsible for designing and owning the probabilistic modeling engine at the core of our

data platform. This involves building large-scale simulation models from first principles,

validating statistical outputs against empirical benchmarks, and ensuring all computational

models perform reliably under live production conditions.

Ø Simulation Design: Build and maintain large-scale Monte Carlo simulations on high-

volume, multi-dimensional datasets using hardware-accelerated parallelism and

compiled code for production-level performance.

Ø Statistical Modeling: Develop census-anchored population models using Bayesian

inference and synthetic data generation to represent large, complex real-world

populations at scale.

Ø Model Calibration: Identify and correct for source signal bias and data duplication

across heterogeneous input datasets. Validate outputs through rigorous statistical

testing.

Ø Output Engineering: Deliver model outputs — including score distributions, ranked

results, and scenario analyses — at strict latency requirements within a live SaaS

platform.

Ø Scientific Ownership: Own the modeling methodology end to end from design to

production, maintained as a version-controlled, property-based tested, and publicly

defensible scientific standard.

Essential Skills and Qualifications:

Ø Education: MSc in Statistics, Applied Mathematics, Computer Science, or a closely

related quantitative field.

Ø Production Simulation: Large-scale stochastic models shipped and maintained in

real production environments — not only research notebooks.

Ø Bayesian Modeling: Working knowledge of PyMC, Stan, or a comparable probabilistic

programming framework used in production.

Ø Synthetic Data Generation: Experience producing statistically valid synthetic datasets

at scale using CTGAN, VAE-based methods, or similar tools.

Ø Scientific Python: Strong proficiency with NumPy, SciPy, JAX, Numba, and

performance profiling tools as part of a daily engineering workflow.

Ø Stack: JAX · Numba · PyMC 5 · CTGAN / Gretel-Synthetics · Scalene + Memray ·

• MLflow — no perfect match required.