AI Engineer YouTube · May 20, 2026

From 46% to 90%: Fine-Tuning Tiny LLMs for On-Device Agents — Cormac Brick, Google

From 46% to 90%: Fine-Tuning Tiny LLMs for On-Device Agents — Cormac Brick, Google video thumbnail
Why it matters

Function Gemma ships at 270 million parameters and processes nearly 2,000 tokens per second prefill on a Pixel 7. Out of the box, on a fixed set of app intents, it hits 46% accuracy. Fine-tuned on a synthetically generated dataset, it clears 90% on eight of ten functions. Cormac Brick covers the two options developers

My takeaway: Function Gemma ships at 270 million parameters and processes nearly 2,000 tokens per second prefill on a Pixel 7. Out of the box, on a fixed set of app intents, it hits 46% accuracy. Fine-tuned on a synthetically generated dataset, it clears 90% on eight of ten functions. Cormac Brick covers the two options developers