Skip to content
View Mattbusel's full-sized avatar
🎯
Focusing
🎯
Focusing

Block or report Mattbusel

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 supported. This note will be visible to only you.
Report abuse

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

Report abuse
Mattbusel/README.md

Matthew Charles Busel

Building the primitive layer for production AI systems.

642,000+ lines of code across Rust, C++, and Python.

Star History Chart


Projects

Stars Repo What it does
47 tokio-prompt-orchestrator Multi-agent LLM orchestration. Ran 24 simultaneous Claude Code agents. Circuit breaker, dedup, retry, backpressure. Self-improving loop live.
23 Every-Other-Token Intercepts every other token mid-generation, scores confidence, visualizes perplexity in real time. Research tool for LLM interpretability.
14 LLM-Hallucination-Detection-Script Framework-agnostic Python + Rust toolkit for detecting hallucinations in LLM outputs.
7 llm-cpp 26 single-header C++20 libraries for LLM infrastructure. Zero dependencies. Drop in what you need.
5 Reddit-Options-Trader-ROT Financial intelligence platform. 200+ users across 46 countries. Validated signals: MASI +34%, ZIM +25%, OLB +137%.

Rust Primitive Layer

rust-crates — 10 production libraries, all used inside tokio-prompt-orchestrator:

Crate What it does
tokio-agent-memory Episodic, semantic, working memory with decay and multi-agent bus
tokio-memory Ebbinghaus forgetting curve + async persistence
llm-budget Hard budget enforcement across agent fleets
llm-wasm LLM inference primitives for WebAssembly and edge
llm-diff Semantic diff and versioning for LLM outputs
llm-sync CRDTs and vector clocks for distributed agent state
wasm-agent ReAct agent loop for WASM with resumable state
mem-graph Knowledge graph: typed nodes, edges, BFS/DFS
fin-primitives Price types, order book, OHLCV, indicators, PnL
fin-stream Lock-free SPSC ring buffer, 100K+ ticks/second

Finance + Research

Special-Relativity-in-Financial-Modeling — What if price velocity has a speed limit? Lorentz transforms on OHLCV data. VR=1.27×, Bartlett p=6×10⁻¹⁶. DOI: 10.5281/zenodo.18639919

LLMTokenStreamQuantEngine — C++ engine that maps LLM token semantics to real-time trade signals at sub-second resolution.

12 DOI-indexed papers on Zenodo. Active DARPA ERIS submission.


Stack

Rust / C++ / Python / Tokio / WebAssembly / CRDT / MCP

New York / linkedin.com/in/matthewbusel / mattbusel@gmail.com / medium.com/@mattbusel

Total Stars

Pinned Loading

  1. tokio-prompt-orchestrator tokio-prompt-orchestrator Public

    Multi-core, Tokio-native orchestration for LLM pipelines.

    Rust 49 4

  2. Every-Other-Token Every-Other-Token Public

    A real-time LLM stream interceptor for token-level interaction research

    Rust 23 1

  3. llm-cpp llm-cpp Public

    The C++ LLM toolkit. 26 single-header libraries for streaming, caching, cost estimation, retry, and structured output. Drop in what you need.

    7 1

  4. Reddit-Options-Trader-ROT- Reddit-Options-Trader-ROT- Public

    A modular research pipeline that turns trending Reddit discussions into structured market events and options trade ideas.

    Python 5 2

  5. rust-crates rust-crates Public

    Production Rust libraries for AI agents and financial systems — zero panics, typed errors, aerospace-grade test coverage

    4

  6. LLMTokenStreamQuantEngine LLMTokenStreamQuantEngine Public

    A low-latency, C++-based simulation engine that ingests token streams from an LLM in real-time, maps semantic token meaning to trade signals, and triggers micro-adjustments to a trading algorithm o…

    C++ 2