Snowflake logo

Inspiration

I've been nerding out with RAG applications for a while now. The coolest thing I discovered? Uploading content from my favorite creators and having real conversations with their knowledge base. It's wild - instead of watching 10 hours of videos to find one specific answer, you get exactly what you need

When this hackathon dropped, with Snowflake's tech stack and New Year's resolution season incoming, I knew it was time to turn this side workflow into a product

What it does

Think of it as having a health expert on speed dial - all trained from long form content. Ask questions in the chat UI - and then hit one button and turn that chat into a personalized protocol. No more generic advice or endless content rabbit holes - just clear, actionable guidance when you need it.

How we built it

Snowflake Cortex Search doing the heavy lifting on content retrieval Mistral LLM for digesting and carrying the conversation part Streamlit making everything look clean and made for an easy deployment

Challenges we ran into

The AI part wasn't even the hardest bit. Streamlit within Snowflake made that surprisingly straightforward.

The real head-scratcher was getting everything working locally so I could deploy on Streamlit Community Cloud. What worked perfectly in Snowflake's environment needed quite a bit of trial and error to get working on my machine. We got there in the end.

Accomplishments that we're proud of

I built something that actually helps people cut through the noise. Instead of drowning in content or getting generic ChatGPT responses, users get specific, evidence-backed guidance pulled straight from expert sources. I believe use cases like this will grow more and more in the coming years.

What we learned

The tools out there are getting ridiculously good. You can build stuff in days that would've taken months before.

What's next for Snow Coach

  • Expanding our expert content library (knowledge is power 🧠)
  • Building structured 8-week transformation programs
  • Adding progress tracking to keep people moving forward

Built With

  • cortex
  • mistral
  • python
  • snowflake
  • streamlit
Share this project:

Updates