SemiAnalysis x Fluidstack Hackathon
SemiAnalysis x Fluidstack is kicking off GTC with Power to Prefill, Dirt to Decode, Transformers to Transformers: A Full-Stack AI Infrastructure Hackathon.
This is a 48-hour hackathon that spans the entire AI infrastructure stack and we mean the entire stack. Most hackathons start at the API layer. This one starts at the dirt.
Whether you're building tools for land permitting, optimizing power delivery, automating hardware procurement, orchestrating GPU clusters on SLURM, fine-tuning models, or shipping AI applications, there's a place for you here.
The next wave of breakthroughs won't come just from better models, but from reimagining every link in the chain that turns raw land into decoded tokens, megawatts into prefill, and electrical transformers into transformer models.
Civil engineers, systems programmers, ML engineers, and everyone in between are welcome. Fluidstack will issue compute grants, including Claude API credits and managed GPU cluster access, so you can build something real, solo or as a team.
Come build at the layer you know best, or the one you've always wanted to understand. Register via Luma to be considered, approvals will be granted accordingly.
Note: Luma approval is required for attendance of this event. QR Codes will be scanned.
-------------------------
👀 Win an Exclusive Hardware Prize from NVIDIA
🚀 A DGX Spark for First, Second, Third place.
🚀 Credits from Thinking Machines, Cursor, Morph, Cognition, Windsurf
🚀 One year of Cursor Ultra
More amazing prizes to be announced...!
-------------------------
Agenda Highlights:
Friday March 13 (Virtual):
5pm Kick-off Virtual Webinar (Optional)
5pm Compute Goes Live
Saturday March 14 (Virtual):
12:01am - 11:59pm Use Your Compute All Day
Sunday March 15 (In-Person):
10:00am Check In, Breakfast
10:15am SemiAnalysis Kickoff
10:30am-11:15am Keynotes Begin
11:15am-11:30am Fluidstack Closing Keynote
11:30am Hacking begins!
12pm Lunch
5pm Judging and awards
6pm Closing Remarks, Happy Hour Starts
7pm Event Close
Speakers:
Dylan Patel, Jordan Nanos - SemiAnalysis
Gary Wu - Fluidstack
Thomas Raoux - OpenAI
Mark Saroufim - GPU Mode
Horace He - Thinking Machines