#on-device-ai

Microsoft AI brings MAI-Code-1.1-Flash to GitHub Copilot for local execution

Developers will be able to run agentic coding tasks on-device using MAI-Code-1.1-Flash inside GitHub Copilot without paying inference fees for local model calls.

EmbeddingGemma 2 operates on 191MB to 567MB of RAM with an 8K context window

Google detailed operational specifications for EmbeddingGemma 2, stating the 740M model uses between ~191MB and 567MB of active RAM, expands its context window to 8K, and processes up to 5.5 minutes of audio, 29 images, or 58 video frames per pass.

Google for Developers details offline execution and modular deployment for EmbeddingGemma 2

EmbeddingGemma 2 runs multimodal workloads entirely offline with zero server calls. The model features a modular design that lets developers drop unused vision or audio components to save memory, offers flexible dimension sizes that reduce local database storage by up to 6x, and pairs with Gemma 4 for efficient on-device RAG pipelines.

DeepLearning.AI offers free course on building on-device AI memory

Building AI Assistants with On-Device Memory covers a local system for text, voice, and images and is taught with Qdrant. The course includes a new AI coding lab, and enrollment is free.