Data Scientist

What you'll do

• Build and maintain projection models for player props, spreads, totals, and moneylines.
• Own calibration and backtesting, and publish the accuracy numbers the product shows users.
• Work from a historical store of box scores and odds snapshots in Postgres and S3, and ship the features your models need.
• Find edge by comparing model output against market lines, and quantify it honestly.

What we're looking for

• Strong applied statistics: probability calibration, Bayesian methods, and time-series validation.
• Fluent Python with the modeling stack (pandas, scikit-learn, and a gradient-boosting library).
• Confident SQL against Postgres, and comfort reading large Parquet datasets straight out of S3.
• Experience evaluating models against a moving benchmark rather than a static holdout set.
• The instinct to distrust a result that looks too good — and the rigor to find out why it does.

Nice to have

• You have modeled a betting market, or another market where the price already contains most of the information.
• Experience turning a model's uncertainty into something a non-technical user can read and trust.
• Published or open-source work on calibration or forecast evaluation.

What you'll work with