← All products
MindActive since Sep 2026

SLM Build for Team

Teaching Kit · Hybrid RAG

What it offers

A live, hands-on teaching kit — it builds a real hybrid-retrieval RAG pipeline and fine-tunes a real model in front of a room, with every layer, from chunking and embedding to retrieval and training, genuinely inspectable rather than asserted.

Teaching RAG and fine-tuning usually means a slide deck or a black box — this makes every layer a literal, clickable, checkable step.

SLM Build for Team

How it started · 4 Sep 2026

“PROJECT: "SLM Build for Team" — a live, hands-on tool that builds a RAG and fine-tunes a model from the ground up, step by step, teaching every layer. Doubles as a reusable kit the team points at their own docs. Read...”

1 session later → SLM Build for Team

How it evolved

  • Built as a live, presenter-run teaching session
  • Completed a real fine-tuning run, exported a working model
  • First live comparison surfaced an honest, unscripted finding
  • Runs on two local processes — no cloud model needed

Patterns

1session
~20Kwords
~27Ktokens
Building · 1

From Shape — the shape of the conversation, not its content.

AI use · 1 session

This page shows the shape of the work, not the how. Want to see it running, or talk through how it's built?

Get in touch →