Meta Muse Code | NFL Big Data Bowl 2027

If you've been out towards The Edge of Texas Steakhouse lately, you've noticed Meta is well into building a $10B+ AI data center in El Paso. Having the hardware in our backyard added is intense. Fortunately we do get benefits in the terms of access to the tools Meta is creating with model architectures like the Muse family and coding harness tools like Meta Muse Code under the division of New Mexico born Chief AI Officer Alexandr Wang.

Meta Muse Code is Meta's new CLI software which was released very recently, and after a month with it's power usage ($50/month) access to Muse Spark Contributor and free-to-use local Glimmer models, we are opening the terminal to run it live. We'll discuss Muse Code's remarkable strengths like backgrounding and fleet dispatch. We'll also check out the relative affordability in powering workloads 10x the cost elsewhere.

We'll dive into Muse Code through the lens of the NFL Big Data Bowl, which historically unveils its challenge the last week of September. We'll take a look back at years past to try and gain competitive advantage heading into this year's challenge.

The NFL Big Data Bowl gives the public rare access to the NFL's raw game data in formats like 10Hz player-tracking coordinates (x, y, speed, orientation). Each year the NFL asks data scientists to quantify asks like where receivers and defenders will land while the passed football is in the air. Past winners like Robyn Ritchie (2022) built physics-based return paths, and Philipp Singer (2020) built expected rushing yard metrics that now run on live broadcast graphics.

No Machine Learning, programming, or football background needed. Principles apply to other sports science and human movement domains. As NFL Network judge Cynthia Frelund puts it, the goal isn't box-score statistics—it's process over outcome. If you can read a scatterplot, you can build a model, compete, and build a career.

Open mic- questions, distractions & tangents welcome.

Bring a laptop and code along, or just grab a drink, watch the terminal, and steal the prompts.

Esta sesión será en inglés. Si quieres apoyar como voluntario con traducción al español o LSE para futuras sesiones, escríbele a Bryan 👐

Amazon Wev Services (AWS) Sample - High Speed Ball Tracking
https://github.com/aws-samples/aws-data-driven-sports