Skip to content
View nicholasspencer's full-sized avatar
  • Austin, TX
  • 21:35 (UTC -05:00)

Sponsoring

@rrousselGit

Highlights

  • Pro

Organizations

@Smashtron @resideo

Block or report nicholasspencer

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
nicholasspencer/README.md

Howdy y'all

I'm Nico — Lead Software Engineer at @Resideo, based in Austin.

Been building software for over twenty years now. Started with PHP and MySQL back in '05, picked up iOS development early on, and it stuck. Spent time at Match, Shyp, Synack, and American Airlines before landing at Resideo, where I lead mobile engineering. Most of my work these days is in Flutter.

What I'm tinkering with

Multi-agent orchestration (the factory loop)

Most of my weekend hacking lately is on autonomous agent workflows — a pipeline where AI agents pick up work items (called beads), write blueprints, implement code, review each other's PRs, and land changes with minimal human input. A supervisor coordinates the whole thing: assigns priorities, resolves blockers, and keeps the factory moving.

Some live numbers from the hobby bench:

  • 65 beads tracked across projects
  • ~88% completion rate (57 closed)
  • 310 commits landed across factory-managed repos
  • 5 unique contributors (a mix of human + agent identities)

Built on top of OpenClaw with custom orchestration layers. Still very much weekend-science-experiment energy, but the agents are getting surprisingly productive.

Local inference on Apple Silicon

Also building a local inference orchestration server in Swift — runs multiple model backends (MLX, FlashMoE) behind a single OpenAI-compatible API and routes requests to whichever node makes sense. Early telemetry from real requests on an M-series Mac:

Metric Value
Avg time-to-first-token 331 ms (p50: 117 ms)
Avg throughput 26.7 tokens/sec
Peak local throughput ~65 tokens/sec (7B model on MLX)
Backends tested MLX (7B, 32B), FlashMoE (397B)

Still early days — the goal is to publish reproducible benchmarks for different model sizes on consumer Apple hardware. Live metrics dashboard: claw.nicospencer.com/metrics

🤖 - OpenClaw


Last updated: March 29, 2026

Pinned Loading

  1. dart_optics dart_optics Public

    Lenses, Prisms and other optics for Dart

    Dart 1

  2. .gitconfigs .gitconfigs Public

    Common git settings

    1

  3. gh-skill gh-skill Public

    A gh extension for managing AI agent skills as GitHub Gists

    Go 1