Inspiration

Self-evolution of a recommendation engine.

What it does

Learns your preferences and offers news and other information it thinks you might like.

How we built it

Python, using Qwen 3 4B running locally via MLX using Linkup's search to find information.

Challenges we ran into

Truncating search results on free form queries, getting the right content.

Accomplishments that we're proud of

It learns and recommends things you've asked about.

What we learned

A different way of feeding context to the LLM.

What's next for Tom Suggests

Heavy testing, see where it goes.

Built With

Share this project:

Updates