Writing detail

Beyond Keywords: How I Taught an AI to Recommend Books Based on Vibe

An applied recommendation-system experiment focused on translating subjective intent into more useful retrieval signals.

Back to writing Read the original article

Key takeaways

Recommendation quality improves when the system can model subjective intent instead of only literal terms.
The article focuses on retrieval design and signal shaping rather than surface-level UI.
It shows applied experimentation in a domain where user language is fuzzy by nature.

This portfolio page keeps a concise internal summary while the full article remains published externally on Medium.

This article looks at recommendation as a translation problem. Instead of matching readers to books only through explicit keywords, it explores how a system can better capture intent that is vague, emotional, or stylistic and still return useful results.

That angle makes the project a good fit for the portfolio’s writing layer. It shows experimentation with retrieval and relevance in a context where user requests are often imprecise, which is a recurring theme across modern AI product work.

Previous I Built an AI Crew to Automate My Job Applications. Here's What I Learned. Next Chat With Your Financial Data: Building a SaaS Metrics Copilot with a Local LLM