
Engram: Cognitive-Inspired AI Memory Architecture
Building a production memory system for conversational AI based on cognitive science principles. Dual storage, temporal validity, uncertainty tracking, and entity deduplication in production.
Insights and perspectives on AI, digital health, and emerging technologies

Building a production memory system for conversational AI based on cognitive science principles. Dual storage, temporal validity, uncertainty tracking, and entity deduplication in production.

Evolutionary search over AI agent orchestration patterns. 18 experiments across 6 models (7B–405B parameters) reveal that orchestration effectiveness depends entirely on model scale.

Building a production voice assistant with local GLM routing and semantic memory. 85% of queries handled at 700ms latency with smart home integration and Engram memory.

Porting Kyutai's Mimi neural audio codec to Apple Silicon using MLX. On-device speech encoding and decoding with no cloud dependency, running at native Metal GPU speeds.