JevMade field notes / Retrieval evaluation walkthrough

A guide to building a movie recommender with AI choices

This guide explains how to build a movie recommendation system. It shows how to use basic search to find a shortlist of films, and then use an AI tool to pick the best match based on user rules.

Original by Ariel BubisRetrieval

Listen to this guide

JevMade’s plain-English explanation

0:00 /

AI narration

Credits and license

“Jevflix” by Ariel Bubis. Read the original source.

This expanded guide is an AI-narrated adaptation of the source’s essential explanation, examples and caveats, not a word-for-word reading. The synthetic voice does not imitate the author or imply their endorsement.

License: MIT

MIT License

Copyright (c) 2026 Ariel Bubis

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.

Our summary

This project is a guide for building a movie recommendation system. A person would use this setup to find films based on specific requests, like asking for a short science fiction movie without graphic violence, or by listing other movies they already enjoy.

The system first searches a catalog of films to create a shortlist. It then sends this list to Jev, an AI tool that chooses from options rather than writing an answer. Jev picks the best film and gives a confidence score to decide if it should ask a follow-up question.

This guide is useful for people learning how to test different parts of an AI assistant separately. The author tested the system with specific question sets, but these results happened in their own tests and have not been independently verified by others.

Key takeaways

Search for a shortlist of items before asking the AI tool to make a final choice.
Set clear rules for what the software should do if it is not confident in its answer.
Test the search step and the decision step separately to see which part improves the results.

More guides →See the experiments →