AI & Agent Evaluation
2,152total visitsadmin
reading room / notes / ai engineering

Reading room

Short summaries of agent optimization and AI engineering work — harnesses, prompt and context optimization, orchestration, memory, and agentic RL.

$ evals.index --public
category: ai engineering
posts: 1
mode: short summaries
storage: postgres
status: listening
ai eval (43)ai engineering (28)
Sort by source dateLatest firstEarliest first

Filtering by agent memory. Clear filter.

MemCompiler: Compile, Don't Inject — State-Conditioned Memory for Embodied Agents

arXiv paper · source date 2026-05-01 · 0 comments · original
agent memorycontext engineeringreliability

1. Problems / challenges / motivations - Retrieval-style memory injects a blob of past experience ahead of time, most of which is irrelevant to the current step. - Irrelevant memory costs tokens, adds latency, and actively distracts the policy. - Vector recall quality is not the same thing as memory being useful to an acting agent. 2. Key ideas - Replaces...