Victor Savkin on Agent Memory Across Repositories

AI Engineer20m
Victor Savkin discussing agent memory.
0 comments · 0 votesOpen discussionClose discussion
Sign in to join the discussion

    Victor Savkin argues that repository boundaries and lost session history force developers to repeat context to coding agents. He illustrates the problem with a change that must pass through multiple repositories.

    He presents Polygraph, an agent-agnostic harnessAn AI agent harness is the software framework that packages a model with tools, instructions, context management, execution controls, and user interaction. that builds a dependency graph, coordinates changes and CI across repositories, and retains agent tracesAI agent observability makes an agent's state, actions, tool use, failures, resource use, and outcomes visible enough to understand and operate it..

    In demonstrations, Victor Savkin resumes a shared session on another machine, asks an agent to investigate a previous change and finds related work through the graphRetrieval is the process of selecting relevant stored information and returning it to an AI system for the current task..

    Original YouTube thumbnailWatch on YouTube

    Video topic coverage

    • Agents & workflows100
    • Coding & tools100
    • Models & learning0
    • Infrastructure & compute65
    • Research & evaluation25
    • Safety & security0
    • Business & productivity65
    • Creative & media0
    • Robotics & perception0
    • Society & policy0

    Share this video

    Victor Savkin discussing agent memory. With WWW.ARTIFICIAL-INTELLIGENCE.VIDEO, 26 June 2026 and duration 20m.

    Victor Savkin presents a repository graph and shared session history for agents working across codebases.