I build AI systems and help developers understand the engineering behind them.
I'm a senior engineering leader, AI architect, founder of AgenticWorks, and a Google Cloud Next ’26 speaker. My work spans retail, payments, and intelligent automation; my public projects and teaching focus on Google Cloud, Gemini, ADK, and agent evaluation.
Learn · Talk and team · Community contributions · LinkedIn
| You want to… | Start with… |
|---|---|
| Understand neural networks from first principles | NeuralNet Fundamentals: a guided Python and NumPy track with notebook source. |
| Understand how ADK agents fit together | My illustrated ADK guide: tools, orchestration, state, evaluation, and deployment. |
| Explore a team-built Cloud AI application | AI-mmunity: a community-health agent prototype using ADK, Gemini, BigQuery, and Cloud Run. |
I presented Silos to Synergy: Architecting Scalable Multi-Agent Systems at the Developer Theater. Our AI-mmunity team won Google Cloud's Agentic AI Arena.
I led the team, coordinated and integrated teammates' code, and set up agent orchestration. The talk explored data readiness, evaluating the whole agent system, and managing permission boundaries.
Project and team · Architecture and lessons · Technical recap and discussion
| Resource | What it offers |
|---|---|
| AI Trainings | A search result is not an answer: inspect PDF evidence locally, then explore three Gemini, vector-search, and RAG labs. |
| CX Lab | Reproduce a policy evaluation: compare a baseline, inspect failures, and apply safety checks in a synthetic exercise. |
| Neural Network Fundamentals | Ten notebooks behind the AgenticWorks learning track. |
- Agents & models: Gemini · Vertex AI · Google ADK (labs, guide).
- Cloud & data: Cloud Run · BigQuery · Firebase (AI-mmunity).
- Research & teaching: Python · NumPy · Jupyter (notebooks).
How can AI support learning and social connection? My research with Georgia Tech's Design & Intelligence Laboratory and the National AI Institute for Adult Learning and Online Education (AI-ALOE) explores socially aware AI, learner engagement, and social presence in online classrooms.
What happens when language models interact? Through LLM Arena, I investigate cooperation, competition, persuasion, and adherence to assigned goals. The public code supports configurable experiments and judge-based evaluation.
Education: M.S. in Computer Science, Georgia Tech · M.S. in Applied Artificial Intelligence, University of San Diego.
- ADK as a software-engineering lifecycle: an illustrated guide with an order-operations example.
- Lessons from recent AI projects: reflections on agent improvement, grounding, and interface design.
I'm an active participant in Bay Area hackathons, with multiple team wins. I also enjoy hosting hackathons with companies and local developer communities. Building alongside other developers is one of my favorite ways to learn and share ideas.
Try a lesson, share where you got stuck, or suggest a clearer example. Questions and corrections are welcome in the relevant repository's issues or the AgenticWorks forum.
Explore my teaching, talks, and public contributions →
Inventions
I am a co-inventor on patents for real-time product interaction assistance: US11106327B2 and CN112771472B.