Judging the AITX Hackathon in Austin, Texas!
What an honor to mentor & judge Austin's first AI Hackathon alongside brilliant minds from Antler, Google, Lambda, Sindarin - Conversational Speech AI, and Finetune!
From the Antler HQ in Austin, the teams hacked together some cool projects on Saturday including:
- Graph RAG for podcast content with Neo4j (overall winner)
- AI enabled browser assistant for the visually impaired via Chrome extension (best "vision" project)
- A multi-agent framework for customer persona development (best Austin-based tech stack project)
My personal favorite was from a team utilizing Unity's AI packages to spawn 2 agents into a 3D arena where they would battle to victory. I don't think the team realizes how important this kind of solution is for the coming era of robotics. Being able to create synthetic training data from virtual environment interactions is really the only way we can produce generally capable robots, given the vast amount of training data needed from computer vision sources, environmental data, etc. Reminds me of what Jim Fan's robot team is working on at NVIDIA!
I also had a little time at the end of the competition to submit my own hack ;). I put the finishing touches on my Apple Shortcut tool called FUTURESELF. It pulls in my to-do list from Apple Reminders, feeds that into Ollama locally to draft a motivational message from my to-dos, then it uses ElevenLabs to read that written message in my own voice (cloned from a ~1min recording of me talking). Ultimately, users are texted an MP3 of their own voice motivating them to complete their to-do list.
Shoutout to Jake O'Shea, Achyut Sarma Boggaram, Michael Daigler and the AITX & AI Tinkerers groups for bringing together such a brilliant group of hackers from across Texas. Can't wait for the next hackathon!
