Character & Simulation TA / Applied-AI Engineer
For 25 years I've built digital-character systems for film and AAA games, from The Matrix to World of Warcraft. I now build applied-AI agents and the evaluation harnesses that test them.
I was Technical Lead of the Avatar Hair team at Meta, Rigging & Simulation Lead for seven years at Blizzard, and lead of a CharacterFX team of 25+ at DreamWorks. The work was systems rather than shots: one rig or solver that has to hold up across a whole slate of characters.
A cloth sim can read perfectly in one frame and be broken in every other. A result can look right and still be wrong. Generative and agentic systems fail the same way. So I build agent systems and retrieval pipelines together with the evaluation harnesses that prove they still work.
At Meta I built a multi-agent deep research system, taking it from prototype to a company-shared tool adopted by more than 50 peers, then measured, fixed and re-scored it across 200+ ranked versions against a 15-axis research-performance rubric. I shipped devserver loops that rewrote their own skills from error-log feedback, and shared-cloud frameworks for engineers and OpenClaw. The same gating applies to assets: a C++ FBX to OpenUSD library that validates and auto-corrects non-compliance, shipped as a CI gate so bad geometry could not reach the engine, across hundreds of thousands of assets.
An independently authored LiveKit realtime runtime for Hermes Agent. Speech answers in milliseconds beside agent work that runs for minutes, each with its own cancellation scope, so interrupting a sentence never kills the task behind it. Dispatch fails closed without a real delegation handle. Controls include stop by exact handle and request-bound approvals. Experimental.
Architecture noteAn Unreal Engine 5.8 skeleton-agnostic Animation Blueprint Template for varied rigs. Each graph is golden-diffed and asserted per commit against a reviewed baseline, with live per-node blend-weight introspection. Authoring uses open-source monolith.
Architecture notemainframe-mcp is a 100% local GPU knowledge server using Qwen3 embeddings with a native Qwen3 reranker, hybrid search, and conflict-aware ambient memory. Its eval harness measured +26% composite and +33% hit@1 against bge-reranker-v2-m3.
Architecture noteclaude-oracle fans one question across up to eight chains of ten Haiku scouts, eighty in flight, each with a Sonnet compressor that sees only its own chain. Cheap models do the breadth, a compressor per chain does the judgment, and the Opus session makes the editorial call. Eighty scouts read wider than eight Opus agents and cost less per research dimension covered. MIT-licensed.
Architecture notevram-mcp shares one NVIDIA GPU through TTL claims, NVML detection and protected eviction. hardline-mcp provides Claude Code, Codex and Hermes agents a durable mailbox.
Architecture noteA FLUX.2 LoRA generates a 94-character typeface from two glyphs, vectorized into Bézier outlines for an installable OTF and WOFF2. I built one metric in its four-metric harness.
Architecture noteSports forecasting over roughly 7 million rows in 32 tables. Event-level shots and matchups feed 39 LightGBM quantile models. League-specific paths and walk-forward validation prevent leakage. A PuLP MILP solver selects constrained lineups.
Architecture noteNatural language and voice control for classic adventure games. Clean-room AGI and Z-Machine interpreters expose live state. Rule-map grounding prevents invented commands. Autopilot executes validated actions with dead-state protection.
Architecture notegray.bot uses a database-free content pipeline with automated checks and human approval. TypeScript and Sharp pixel analysis find gray images. An LLM drafts posts for an Astro site on Cloudflare Pages. The archive remains small.
Architecture notePublic repositories are on GitHub. Private demonstrations are available. Meta agentic pipelines and physics-ML training data remain NDA-bound.
Real-time avatar systems
Avatars & Horizon Meta Reality Labs · 2022–2026
Built core avatar rigging and deformation systems in a VR/mobile engine: auto-rigging, autofit deformation, skinning, blendshapes, skeleton standardization, node-graph clothing-rig automation, and runtime animation & simulation graphs within performance budgets. Avatar Hair Technical Lead. Delivered 100 character rigs for a physics-based ML rigging model.
Rigging & Simulation Lead
Blizzard cinematics Blizzard Entertainment · 2015–2022
Full-body, facial, muscle, clothing and hair rigging across World of Warcraft, Overwatch, Diablo, Heroes of the Storm and Hearthstone. Architected a simulation branch of the studio’s modular graph-based rig-build system with PyQt tooling. Led rigging and shot simulation on 25+ cinematics.
CharacterFX Lead
Eight feature animations DreamWorks Animation · 2005–2015
Led a CharacterFX team of 25+ and built hero-character cloth, hair and skin across How to Train Your Dragon, Kung Fu Panda, Puss in Boots, Bee Movie, Over the Hedge, Turbo, Penguins of Madagascar and Home. Developed cloth-based skin solvers and core departmental Python tools.
Autonomous crowd agents
The Chronicles of Narnia Rhythm & Hues Studios · 2004–2005
Motion editing, procedural retargeting for autonomous crowd agents, and motion-capture TD work on The Lion, the Witch and the Wardrobe.
Digital-human facial capture
The Matrix sequels ESC Entertainment · 2002–2004
Facial rigging, sculpting, match-move and texture projection on award-winning Universal Capture for Reloaded and Revolutions. Team member and primary contributor to “Making of The Superpunch”, screened at SIGGRAPH 2004 Electronic Theater.