Smaller, faster, smarter: Distilling models with fine‑tuning | DEM322

Name: Smaller, faster, smarter: Distilling models with fine‑tuning | DEM322
Uploaded: 2026-06-04T10:54:06.000Z
Channel: Microsoft Developer

Model distillation Fine Tuning Microsoft foundry Small language models AI optimization Production machine learning Model compression Cost optimization ML deployment Student model training

Microsoft Developer June 4, 2026

AI summary

This demo shows how to use Microsoft Foundry for model distillation and supervised fine-tuning to create smaller, faster, cheaper language models for production. It covers when distillation is appropriate, how it complements fine-tuning and RL, and includes real case studies from production teams deploying distilled models at scale. Ideal for ML engineers and production teams looking to optimize AI inference costs.