Skip to content
View marklubin's full-sized avatar

Block or report marklubin

Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
marklubin/README.md

Mark Lubin

I build infrastructure for AI agents: runtimes and orchestration, memory and context systems, evaluation, and the backend systems that keep them reliable.

I spent ten years at Amazon building distributed systems across payments, Prime Video, and AWS Marketplace. I now work across product definition and agent-platform infrastructure at Kinelo. My independent research and open-source work produced Synix, LENS, and CDR.

Selected work

  • Synix (source): an open-source agent-memory build system with DAG execution, content-addressed artifacts, incremental rebuilds, and source-level provenance.
  • LENS: a benchmark for long-horizon agent coherence under context accumulation and drift.
  • Compress, Divide, Renew: an architectural framework for coherence in persistent agent systems.
  • What Memory Is (And Isn't): a talk on memory and context infrastructure for agents.

San Francisco, CA · marklubin.me · LinkedIn

Pinned Loading

  1. kairix kairix Public

    Voice-first AI agent with persistent memory, background reflection, and hierarchical summarization.

    Python 11

  2. synix synix Public

    Memory architectures that evolve with your agent

    Python 12 3

  3. lens-benchmark lens-benchmark Public

    LENS - AI Memory Benchmark - Memory as Experience, Not Facts

    HTML 1