Skip to content
View jameswniu's full-sized avatar

Highlights

  • Pro

Block or report jameswniu

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 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
jameswniu/README.md

James W. Niu

Sr. Staff AI/ML Engineer, 10+ years. I build multi-agent systems that ship, and I measure them before I believe them.

Built work, with receipts. Every project below runs, and every claim traces to an eval, a probe, or a logged run. The receipts that proved me wrong are here too.

A wrong number in a dashboard fails loudly. A generative system fails plausibly: a mouth trailing the audio by four frames, a cited statistic with nothing behind it. Those are invisible to a type check, obvious to a person, and different again tomorrow. So I capture human judgment as labels, compile it into thresholds, and wire it into gates that can refuse to spend.

if you want go to
one project, read deeply the featured pipeline
range across problems the table of five
how I work principles I work by

Featured: AI filmmaking governed by multimodal evals

3d-filmmaking-ads-multimodal-evals

13 probes · 16 of 16 gating thresholds derived from labelled exemplars · 4 guards, 3 of which fail open and say so · 1 credit per scheduled render · 77 views per frame

An AI filmmaking pipeline for advertising-grade avatar video: multimodal evals govern a cloned voice, a consistent generated character, and light-field 3D output. Taste is captured as labels, compiled into thresholds, and enforced by gates that block bad output before it spends money.

It runs on a schedule, against metered vendor APIs, with nobody awake. That is the design constraint: an unattended pipeline cannot be corrected mid flight, so every gate fires before the credit is spent, and anything short of a clean success pages loudly instead of passing quietly.

The repo publishes the measurements that falsified my own claims, including two failures the shipped render still does not pass, and a speedup whose number held while the explanation I wrote for it turned out to be wrong.

Five more: Production AI on real problems

Ranked by priority, one per repository, nothing listed twice.

Focus Project What it solves Who benefits
Voice AI
Conversational AI Realtime ASR + TTS Twilio PSTN LLM‑as‑Judge Evals
realtime-voice-agent-turn-taking-stack Answering a real phone line with no human on it, and knowing when to stop talking Support, scheduling, and anyone shipping voice agents
Reinforcement Learning
Multi‑Agent RL PPO PyTorch FastAPI
multi-agent-rl-mapf-drone-navigation Many drones planning conflict free paths at once, in real time Robotics, drone, and simulation teams
LLM Evaluation
Prompt Engineering Self‑Hosted Llama Ablation Benchmarks Statistical Analysis
self-hosted-llm-evals-lab Knowing which prompt strategy actually wins, by statistics instead of vibes ML platform and LLM eval teams
RAG Pipelines
LLM Orchestration Claude + GPT APIs NLP Extraction Document Processing
research_doc_extraction_rag_agent Thousands of open ended survey answers nobody has time to read Research, insights, and product teams
MCP Servers
Model Context Protocol Tool Design OAuth 2.1 Self‑Hosted Atlassian
onprem-bitbucket-jiraconfluence-mcp AI agents that cannot reach self-hosted Bitbucket, Jira, or Confluence Enterprise platform and developer experience teams

Two things I learned by being wrong

A suite only covers what someone thought to measure. Eleven probes scored the face, the motion and the timing in a rendered frame. None asked whether the subject was visible against the background. A dark top on a black matte separated by 22 levels of luma where the face separated by 134, so the torso dissolved and the render shipped a floating head. Every probe passed. The gap was not a bug in any of them; it was a question nobody had written down.

Three green checkmarks over two empty files. In a multi-agent system every step reported success and both artifacts persisted empty. The agent cards advertised localhost, correct on one machine and wrong across a container network, so discovery kept succeeding while delivery failed silently. The client now rebases the address and logs that it did, and every artifact writes its byte size.

Principles I work by

  1. Chat output is not evidence. The database write, the rendered file, and the logged artifact size are the evidence.
  2. Degrade loudly, never silently. Every fallback is logged; a green checkmark over an empty artifact is the worst bug in agentic systems.
  3. Spend where the artifact is judged. Flagship models on deliverables, fast models on hot paths; cost per accepted deliverable beats cost per token.

Skills

  • Languages: Python, TypeScript, Bash
  • LLMs and agents: Multi-agent systems, agentic AI, LangGraph, MCP (Model Context Protocol), A2A (agent-to-agent), RAG, tool design, prompt engineering
  • Evaluation: LLM evaluation, LLM-as-a-judge, guardrails, ablation studies, statistical analysis, regression testing
  • Voice and speech: Conversational AI, voice AI, real-time ASR and TTS, speech recognition, text-to-speech, Twilio, PSTN, WebSocket, voice cloning
  • Machine learning: Reinforcement learning, PPO, PyTorch, NLP, generative AI, computer vision, monocular depth estimation
  • Retrieval: Vector search, embeddings, semantic search, reranking
  • Platform and MLOps: MLOps, FastAPI, Docker, CI/CD, Cloudflare Workers, observability
  • Models: Claude (Anthropic), GPT (OpenAI), Llama

Open to work

Sr. Staff / Principal AI engineering: agentic systems, LLM evals, and the guardrails that make them safe to run unattended. San Francisco, or remote.

Role titles I map to: Staff AI Engineer, Principal AI Engineer, Applied AI Engineer, Machine Learning Engineer, AI/ML Engineer, Forward Deployed Engineer (FDE).

The fastest way to evaluate me is to not take my word for it. No accounts, no API keys, no GPU. Needs python3 and ffmpeg:

git clone https://github.com/jameswniu/3d-filmmaking-ads-multimodal-evals
cd 3d-filmmaking-ads-multimodal-evals && pip install -r requirements.txt
python3 evals/derive.py

It prints every gating threshold in that pipeline beside the labelled pass and the labelled reject that bracket it, recomputes the shipped exemplars rather than reciting them, and exits nonzero if any constant has drifted outside its own evidence.

Hiring, or want the walkthrough: LinkedIn | +1 (917) 355-7504

Pinned Loading

  1. vector-search-rerank-filtering-semantics vector-search-rerank-filtering-semantics Public

    Three-stage retrieval pipeline over a vector database: Voyage-3 embedding retrieval, hard-criteria filtering, then GPT-4o-mini reranking.

    Python

  2. claude-code-in-python-and-prose claude-code-in-python-and-prose Public

    A study of how an agentic coding CLI is put together, told twice: thirty-four prose chapters on the query loop, tool architecture, hooks, sandboxing, and MCP, plus the same architecture expressed i…

    Python

  3. meta-pixel-cookie-tracking-explorer meta-pixel-cookie-tracking-explorer Public

    Interactive exploration of how Meta's tracking pixel and cookie ecosystem works. Four dimensions: perception, structure, source, time.

    HTML

  4. nlp-clinical-trial-eligibility-screening nlp-clinical-trial-eligibility-screening Public

    A modular system for evaluating patient eligibility against clinical trial protocols using structured data and clinical notes. Includes semantic analysis, confidence scoring, and Dockerized deploym…

    Python

  5. personal-whatsapp-actions-api-daemon personal-whatsapp-actions-api-daemon Public

    WhatsApp's official API is Business-only and never touches your personal chats or groups. This wraps the WhatsApp Web protocol in a local HTTP API any client can drive: agents (OpenClaw, Claude Cod…

    Go

  6. questions-to-sql-intent-router questions-to-sql-intent-router Public

    An intent classifier routes plain-English questions to document retrieval (FAISS) or to SQL generated from the question and validated before it runs. 100 percent query validation, PII and prompt-in…

    Python