This past weekend, I had an amazing time attending the AI Coding Agents Hackathon, hosted at Y Combinator by Anthropic, Freestyle Cloud, Same, and Morph — alongside my teammate Aeliya Grover.
In just ~7 hours, we pushed HandsIn, our American Sign Language (ASL) learning platform, further than we thought possible.
What We Built 🚀
HandsIn already helps learners practice ASL using hand tracking and adaptive lessons. For this hackathon, we:
- 🔁 Implemented a Q-learning reinforcement learning algorithm to power a new adaptive flashcard system that personalizes learning based on user performance
- 🖥️ Began training a new model for our upcoming ASL course on basic emergency terms
- 🎨 Kicked off a website redesign to improve the learning experience
- 🎬 Built and demoed a working prototype — all in one day!
What We Learned 💡
Hackathons are always intense, but they’re also a reminder of how much can be accomplished with focus, creativity, and collaboration. We were blown away by how quickly tools from companies like Anthropic, Freestyle Cloud, Same, and Morph helped us prototype and iterate.
A special shoutout to Benjamin Swerdlow for inviting us, and to Tejas Bhakta and Aiden Bai for helping run such a smooth, energizing event.
Next Steps 📈
This was just the beginning. Our immediate goals are to:
- Integrate the new Q-learning flashcard system with our existing HandsIn platform
- Finalize and launch our emergency ASL course for learners on our waitlist
- Continue refining the website and production pipeline for a wider release
If you’d like to join the waitlist for our upcoming ASL courses, sign up here: handsinlearning.co
Conclusion
We left the hackathon exhausted but inspired — grateful for the chance to experiment, build, and connect with such a talented community.
This is just the start. Imagine what we can do with more than 7 hours!