Skip to content

Courses 1 open · 3 courses

AI EngineeringStart with Day 1 · The AI Engineer: Role, Skills & How to Use This Course, 0 of 100 doneStart →
Machine LearningBeing written · opens 2027Coming soon
System DesignBeing written · opens 2027Coming soon
All 100 lessons →

AI Engineering · Phase 10 · Days 91–96

Open Models and Fine Tuning

Advanced: run open-weight models yourself, and fine-tune one when prompting isn't enough.

6 lessonsSee the whole track →

10

Phase 10 · Open Models and Fine Tuning

Days 91–966 lessons0 done

Advanced: run open-weight models yourself, and fine-tune one when prompting isn't enough.

  1. Day 91: Open-Weight Models & Local InferenceOrbit's security reviewer asks a fair question: "Can Atlas answer questions without sending our issue text to a third party?" Right now your R2 assistant calls…35 min
  2. Day 92: Quantization & Hardware SizingThe model you ran on Day 91 loads fine on your 16 GB machine, then a teammate tries it on an 8 GB laptop and it either refuses to load or crawls.38 min
  3. Day 93: Serving Open Models with vLLM (elective)Atlas running on one laptop serves one maintainer well.33 min
  4. Day 94: Should You Fine-Tune? Decision & DataOn Day 92 the small local model scored well below the cloud model on R1 extraction, so "let's fine-tune it" is tempting.39 min
  5. Day 95: Hands-On LoRA Fine-TuningYou have about 600 clean extraction examples from Day 94.36 min
  6. Day 96: Evaluating, Shipping & Distilling Small ModelsOn Day 95, atlas-extract looked great on five smoke-test issues.40 min

← Every phase of AI Engineering