Skip to content
View greynewell's full-sized avatar
78 followers · 40 following

Achievements

Achievement: Pair Extraordinairex4Achievement: StarstruckAchievement: YOLOAchievement: QuickdrawAchievement: Pull Sharkx3

Achievements

Achievement: Pair Extraordinairex4Achievement: StarstruckAchievement: YOLOAchievement: QuickdrawAchievement: Pull Sharkx3

Highlights

  • Pro

Block or report greynewell

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
greynewell/README.md

Grey Newell

ML infrastructure engineer. Founding Engineer @supermodeltools, building code analysis tooling for AI agents. Creator of the MIST stack. MS CS (ML) @ Georgia Tech. Ex-AWS.

MIST stack

MatchSpec · InferMux · SchemaFlux · TokenTrace

Go. Zero external deps. All six repos are pinned below.

Repo Purpose
matchspec Benchmark suites. Runs evals against any backend, produces structured reports.
infermux Inference routing. Abstracts providers, tracks tokens and cost per request.
schemaflux Data compiler. Pass pipeline, pluggable backends, no runtime allocs in hot path.
tokentrace Observability. Collects spans, computes latency percentiles, fires threshold alerts.
mist-go Shared core. Protocol, transport, metrics, circuit breakers, checkpointing.

Methodology: eval-driven development.

Research

ORCID DOI

mcpbr: Benchmarking Model Context Protocol Servers on Software Engineering Tasks — 500-task SWE-bench Verified evaluation showing MCP tools trade resolution rate for efficiency (42% fewer tool calls, 15% lower cost). Georgia Tech, 2026.

Pinned Loading

  1. matchspec matchspec Public

    Eval framework. Define correct, test against it, get results.

    Go 21 8

  2. infermux infermux Public

    Route inference across LLM providers. Track cost per request.

    Go 89 7

  3. tokentrace tokentrace Public

    Where did your tokens go? Spans, latency percentiles, alerts.

    Go 5

  4. mcpbr mcpbr Public

    Forked from supermodeltools/mcpbr

    Model Context Protocol Benchmark Runner

    Python

  5. supermodeltools/mcp supermodeltools/mcp Public

    Supermodel Model Context Protocol server. Generate code graphs in Cursor, Codex or Claude Code!

    TypeScript 4 1

  6. swe-bench-pro-action swe-bench-pro-action Public

    GitHub Action for SWE-bench Pro evaluation powered by mcpbr

    Shell