The Intelligence and Investigations team seeks to rapidly identify and mitigate abuse and strategic risks to ensure a safe online ecosystem.
Requirements
- Expertise with modern toolchains—NumPyro, TensorFlow Probability, PyMC, Darts, GluonTS/Chronos, sktime, PyOD 2.0, River, scikit‑survival—and readiness to evaluate emerging libraries as the field evolves.
- Strong coding skills (Python/JAX/PyTorch or R) and data‑engineering fundamentals (SQL, Spark, data warehousing).
- Deep fluency in statistical inference, forecasting, uncertainty quantification, and decision modeling—especially under sparse or adversarial data conditions.
Responsibilities
- Design probabilistic & Bayesian models using PyMC, NumPyro (JAX‑accelerated HMC/NUTS) and TensorFlow Probability to capture uncertainty at scale.
- Build classical and deep‑learning forecasts with statsmodels baselines, plus state‑of‑the‑art libraries like Darts, GluonTS, Chronos, sktime and Nixtla’s MLForecast for multivariate or long‑horizon time‑series problems.
- Develop real‑time anomaly‑detection pipelines leveraging PyOD 2.0 for GPU‑ready detectors and River for streaming/online ML on telemetry data.
- Apply survival‑analysis and rare‑event methods (e.g., Cox PH, random‑survival‑forests, DeepSurv) via scikit‑survival to model threat lifecycles and hazard rates.
- Run stress tests & Monte Carlo simulations to evaluate the likelihood and impact of low‑frequency, high‑severity threats; translate findings into resilient safety‑engineering requirements.
- Own production pipelines in Python/JAX/PyTorch or R, using SQL or Spark‑like engines (DuckDB, BigQuery, Snowflake) and GPU/TPU acceleration where appropriate.
Other
- 5+ years experience in a quantitative research, forecasting, or risk modeling role in finance, tech, safety, security, or public policy
- Crisp communicator able to influence multidisciplinary partners and executives.
- Comfort navigating imperfect data and prioritizing under uncertainty in a rapidly changing threat landscape.