Skip to content
View jonathan-politzki's full-sized avatar

Block or report jonathan-politzki

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

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

Report abuse
jonathan-politzki/README.md

Jonathan Politzki

Working on computers that understand humans @ Jean Technologies.

Background

My interests in AI started in university, building algorithmic trading systems and founding the University of Illinois' first quant organization. Around the same time I started a biotech ML startup called Nephra, where we trained deep learning models on top of ECG waveforms to classify hyperkalemia (elevated potassium) for patients with end-stage renal disease.

Since 2024 I've gone deeper into programming. Early work grew into what is now called AI memory. See jean-memory, an open-source AI memory layer. I believe we were the first server-hosted memory in the world and later turned the product into an SDK that developers could use for plugging memory into their enterprise applications. For developers who want to learn more about the foundations of AI Memory, I wrote the first comprehensive review of the space in 2025.

Current Work

Memory and context are step 1. Foundation models still can not properly represent humans. I feel the applied AI part of this problem has mostly been solved and I am now doing a bit more fundamental work to solve the more fundamental problems at the core of how we can build these systems.

Most of my public work here is scrappy tests. There is much left to do.

Elsewhere

Writing, contact, and everything else: jonathanpolitzki.com

Pinned Loading

  1. jean-technologies/jean-memory jean-technologies/jean-memory Public

    Forked from mem0ai/mem0

    next-generation AI memory infrastructure (powered by mem0 and graphiti)

    Python 170 13

  2. mcp-writer-substack mcp-writer-substack Public

    Model Context Protocol to bridge in Substack writings to Claude.

    Python 31 6

  3. mixed-adapter mixed-adapter Public

    Taking further the work from DRIFT-adapter

    Python 1

  4. mite-bench mite-bench Public

    MITE: Massive Interaction Task Evaluation — benchmarking embeddings on interaction tasks that MTEB misses

    Python 1

  5. agent-trade agent-trade Public

    Anthropic inspired agentic marketplace

    Python 1

  6. rohankatakam/tracer rohankatakam/tracer Public

    Interviewed with YC in S25

    Python 1