Skip to content
View MaxFreedomPollard's full-sized avatar

Sponsoring

@dtolnay
@ggerganov
@sharkdp
@teknium1

Block or report MaxFreedomPollard

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.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. 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
MaxFreedomPollard/README.md
Max Freedom Pollard

I build tools for AI agents and contribute across open source, with 290+ changes merged into 140+ repositories in Python, Rust, Go, TypeScript, C/C++, and Java for organizations like Nous Research, NVIDIA, Sakana AI, Qwen, Tencent, Hugging Face, Apache, and Mozilla. Most of my focus is on agent memory, evaluation, and self-evolution, including ongoing work on Hermes Agent alongside Hermes Agent Self-Evolution, SkillEvaluator, Molt, ShinkaEvolve, autocontext, and headlong.

Among other projects, I created Compartment, an encrypted, offline memory server for AI agents that ships in the Hermes Agent plugin catalog and has one click setup with Anthropic's Claude Code and Claude Desktop, OpenAI Codex, Cursor, VS Code, OpenClaw, and other MCP clients. Rather than plain vector storage, it keeps single-claim memories with source provenance, links them in an entity graph, tracks opinion updates separately from facts, and recalls them in ~12 ms with hybrid vector and keyword search. I also built Max Watch, a free motion-activated Mac security camera app that records while the screen is locked, and Artificial Knowledge Collection 6.0, a cleaned common-sense and reference knowledge dataset for AI.

Accepted benchmark challenges

Challenges authored for Datacurve's ShipD platform, each accepted after automated checks, solver runs and human review; ten are included here as examples, with the hidden tests and reference solutions withheld to protect Datacurve's proprietary benchmark material:

Merged contributions

Projects

Certifications and education

  • MBA, University of Adelaide (now Adelaide University), 2024
  • Google IT Support Professional Certificate
  • IBM Cybersecurity Analyst Professional Certificate
  • Google Project Management Professional Certificate
  • Library & Information Services, University of Adelaide and the Australian Library and Information Association (ALIA)
  • Foundations of Finance, University of Cambridge
  • Member, Association for the Advancement of Artificial Intelligence (AAAI)
  • Board Approved Member, Council on the Ageing (COTA) NSW, since 2026
  • National Member, National Seniors Australia, since 2025

Contact

linkedin.com/in/maxfreedom

Pinned Loading

  1. Compartment Compartment Public

    Encrypted, fully offline agentic memory. One click install, GUI w/ memory map, all OS and agents. Superior memory creation, storage and retrieval.

    Python 569 8

  2. Max-Watch Max-Watch Public

    Free, 100% offline, encrypted Mac & Windows video recorder with motion detection. Records while your screen is locked into a password-protected AES-256 vault.

    C++

  3. Tencent/BrowserSkill Tencent/BrowserSkill Public

    Let AI agents use your real, logged-in browser without interrupting your work. CLI + extension for browser automation across any shell-capable AI agent.

    TypeScript 10.4k 752

  4. NousResearch/hermes-agent-self-evolution NousResearch/hermes-agent-self-evolution Public

    ⚒ Evolutionary self-improvement for Hermes Agent — optimize skills, prompts, and code using DSPy + GEPA

    Python 5.7k 705

  5. NVIDIA-NeMo/labs-molt NVIDIA-NeMo/labs-molt Public

    A scalable, agentic-first, and HuggingFace-native RL framework for research (9k lines).

    Python 1.3k 139

  6. QwenLM/Qwen-MM-Plugins QwenLM/Qwen-MM-Plugins Public

    Make any agent harness multimodal-native.

    Python 3.2k 227