Skip to content
View rezzminator's full-sized avatar
:electron:
Building aggressively
:electron:
Building aggressively

Block or report rezzminator

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. 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
rezzminator/README.md
A year of contributions drawn as an isometric landscape behind a pane of glass

I build the machinery behind LLM systems — LangChain/LangGraph pipelines, retrieval,
agent tooling, and the distributed services underneath them.

X (formerly Twitter) Stack Overflow Profile views

~ $ whoami

  • 🧠 LLM pipelines — LangChain and LangGraph in Python: stateful multi-step graphs, structured output that actually validates, retrieval over messy real-world documents.
  • 🤖 Agent systems — multi-agent research and QA pipelines on top of Claude Code, with schedulers, claim ledgers, and gates that refuse bad work.
  • 🛠 The infrastructure underneath — Go and Rust for the hot paths, Postgres and vector stores for the memory, Docker and Kubernetes to hold it together.
  • 🧪 Curious to a fault — a WebGL glass shader, an EU-AI-Act validator, a Brainfuck interpreter. If it's interesting, it gets built.
  • 📫 Find me on X.

🧰 Tech I reach for

LangChain LangGraph Claude Pydantic Vertex AI RAG and vector search



Go, Rust, Python, TypeScript, JavaScript, Swift, Docker, Kubernetes, Postgres, Redis, Cassandra, GraphQL, AWS, Linux, Nginx, Bash

📌 Featured work

🎓 professor · Go

A discipline layer for AI coding agents — multi-agent pipelines with schedulers, QA gates, and a pfm CLI that treats every running chat as infrastructure.
Go · Multi-agent · Pipelines

🧩 gemini-coax · Python

Makes Gemini structured output actually validate against your Pydantic models — repairs the anyOf/enum drop and ignored bounds that break LangChain chains.
LangChain · Pydantic · Vertex AI

📊 limit-dashboard · Swift

Native macOS dashboard for AI subscription limits — Claude, Codex and Vertex usage in one live, local-only view.
Swift · SwiftUI · macOS

🌾 harvester-web-mcp · Python

Retrieval front-end for LLM pipelines: web pages, PDFs, Office docs and archives to clean Markdown, with a wall-bypass ladder for walled academic sources.
Python · MCP · Ingestion

📄 resume101 · ⭐ 135

An opinionated resume guide for experienced software engineers. My most-starred project by a wide margin.
Guide · Career

🌫 liquid-glass · WebGL

Zero-dependency liquid-glass effect — smoke refracted through a bevelled glass slab, every parameter a live knob. Live demo »
WebGL · GLSL · Zero-deps



🛰 Right now: professor

Most AI coding setups treat chats as scrollback. Professor treats them as infrastructure — a discipline layer of agent pipelines and QA gates, plus pfm, a Go CLI that manages every AI coding session running on your machine: Claude Code, Codex and OpenCode, across accounts.

 pfm  🥇 account 1 · ⚡1h · 12 rows · 38 killed · 64 empty
 tabs   Chats   Stats   Limits    tab/shift+tab
 Chats · fuzzy search and all existing chat controls
find › type project or name                                                        12/12 visible
╭─ fleet 12 ───────────────────────────────────────────────────────────────────────────────────────╮
│╭─ api                                                                                            │
│› ✦ [ Claude ] Codex OpenCode     🥇                                             0p     0B      0s│
││ ● PAYMENTS_REFACTOR             ⬢ 🥇 ⇄                                       118p    14M      2m│
││ ⚙ SCHEMA_MIGRATION              ⚙ agent 🥈                                    20p   6.1M      7h│
│╭─ webapp                                                                                         │
││ ● ORCHESTRATOR                  🥇 ⇄ ←here                                    37p   2.7M      1m│
││ ⚙ DESIGN_PASS                   ⚙ agent 🥇                                    18p   1.9M     58m│
││ ↻ DEPLOY_PROD                   🥇                                            59p   8.6M      2d│
│╭─ ops                                                                                            │
││ ● FLEET_BUILDER                 ⬢ ⇄                                           77p    94M      0s│
││ ↻ CCC                           🥈                                           452p    32M     54m│
╰──────────────────────────────────────────────────────────────────────────────────────────────────╯
 ↑↓ move  enter open  esc cancel  type to fuzzy-find
 ⌃X hide  ⌃E 1h  ⌃S account  ⌃O reboot

Every AI chat on the machine, grouped by repo — live (●), resumable (↻) or agent-run (⚙), across accounts (🥇🥈). ⇄ marks chats that talk to other chats; ←here is the one you are sitting in. Pick one, attach, or fire it a goal without ever attaching. There is a cosmos tab too, which draws the same fleet as a star map.

📈 GitHub at a glance

GitHub stats: total stars, public repos, contributions, followers, and top languages

🧊 The same year, broken down

3D isometric view of the last year of GitHub contributions

🐍 ...and the snake that eats it

A snake animation eating the GitHub contribution graph

The banner, stats card, 3D graph and snake are generated in this repo by GitHub Actions and committed as static SVGs — the badges are the only rented pixels.

Pinned Loading

  1. resume101 resume101 Public

    Expert resume guide for experienced software engineers

    137 2

  2. professor professor Public

    Multi-agent framework for Claude Code, Codex & OpenCode — coding agents you can watch, message and hold to rules

    Go 8 1

  3. gemini-coax gemini-coax Public

    Make Google Gemini structured output actually validate against your Pydantic models — fixes the anyOf/enum drop, ignored numeric & length bounds, and degraded array tails.

    Python 3

  4. harvester-web-mcp harvester-web-mcp Public

    No longer maintained — this MCP fetch/search server now ships inside the Professor framework as its built-in harvester (fetch, search, findWorks, archive, cache). Use github.com/rezzminator/profess…

    Python 1

  5. limit-dashboard limit-dashboard Public

    Native macOS dashboard for AI subscription limits — Claude/Claude Code usage windows across multiple accounts, OpenAI Codex (ChatGPT) rate limits, and Google Vertex AI token spend. Live reset count…

    Swift 1

  6. liquid-glass liquid-glass Public

    Liquid glass effect for the web — zero-dependency WebGL backdrop: slow white smoke refracted through a thick bevelled glass slab that tracks any element. Glassmorphism in a shader, every parameter …

    JavaScript 1