AP2 — Agent Payment Protocol (Google / google-agentic-commerce)
PR #165 — B2B agent governance framework for multi-round AI-mediated commerce negotiations. Protocol adopted by 60+ Fortune 500 partners. Apache 2.0.
AI architecture across Commerce (BigCommerce, Feedonomics, and Makeswift). Published researcher. 8+ years at Amazon Web Services. Contributor to emerging agentic AI standards.
| Year | Title | Venue | Status |
|---|---|---|---|
| 2026 | Agentic AI for Clinical Urgency Mapping and Queue Optimization in High-Volume Outpatient Departments: A Simulation-Based Evaluation | AICTC 2026 · Springer LNNS | Oral Presentation |
| 2026 | An Agentic Multi-Model Architecture for Cybersecurity Risk Management | AICTC 2026 · Springer LNNS | Accepted |
| 2026 | Pedagogical AI in Mental Health: A Tri-Stream Fine-Tuned LLM Framework for Automated Clinical Supervision and Risk Triage | AICTC 2026 · Springer LNNS | Accepted |
| 2019 | An Explainable Approach of Inferring Potential Medication Effects from Social Media Data | AIME 2019 · Springer LNCS | Published |
| 2017 | Deep Gramulator: Improving Precision in the Classification of Personal Health-Experience Tweets with Deep Learning | IEEE BIBM 2017 | Published |
| 2017 | Identifying Personal Health Experience Tweets with Deep Neural Networks | IEEE EMBC 2017 | Published |
Top ~10% of IEEE's 400,000+ global membership. Requires demonstrated engineering excellence and community impact.
Peer reviewer at the World Congress on Computational Intelligence — IEEE's flagship AI and neural networks conference.
Judge, University of Washington Foster School of Business MBA Case Competition, March 2026.
Recognized for contributions to the general availability launch of Amazon Transform, a major AWS product.
PR #165 — B2B agent governance framework for multi-round AI-mediated commerce negotiations. Protocol adopted by 60+ Fortune 500 partners. Apache 2.0.
First-mover enterprise MCP implementation across BigCommerce, Feedonomics, and Makeswift — defining the foundational agent communication layer for production commerce AI.
Drawn to the problem of model integrity — how AI systems can be compromised silently, at the data layer or the inference layer, in ways that pass every standard quality check. The NIST AI 100-2 taxonomy frames this well: availability breakdown versus integrity violation. I am most interested in the integrity side. Currently reviewing adversarial ML research for IEEE WCCI 2026 IJCNN.
There is a state where thought dissolves into doing. Where the painter, the athlete, the scientist, and the soul are the same person. That state has a name. Krita.
Available for research collaborations, conference speaking, advisory roles, and media inquiries on agentic AI and enterprise AI strategy.