Simulating the Superbowl: Real-Time ML to Predict the NFL (Repeated)
OVERVIEW
EXPERIENCE | In Person |
---|---|
TYPE | Lightning Talk |
TRACK | Data Science and Machine Learning |
INDUSTRY | Media and Entertainment |
TECHNOLOGIES | AI/Machine Learning, MLFlow |
SKILL LEVEL | Advanced |
DURATION | 20 min |
DOWNLOAD SESSION SLIDES |
This talk is repeated.
In the fast-paced world of live sports, anything can happen at any given moment. This creates an exciting challenge for sportsbook operators offering live betting products. How do you quickly react and respond to these events, offering a complete breadth of content and accurate predictions? In this talk, we will showcase how DraftKings does just that, focusing on our football model for the NFL. We will dive deep into our pricing engines, which use Monte Carlo simulations together with an ensemble of machine learning models that predict what might happen in each play. We will also cover our MLOps pipelines, training and deploying our models, and showcase how MLflow and Databricks have helped accelerate our SDLC.
SESSION SPEAKERS
Rohan Shanker
/Senior Lead Data Science Engineer
DraftKings
Niall Creech
/Senior Manager Quants
Draftkings