Building AI-native products that turn messy information into structured insight.

AI Infrastructure Supply Chain Dashboard — Built a Streamlit dashboard to track structural supply-chain bottlenecks across memory, packaging, test, and country concentration using public semiconductor market-share data. Added a priority scoring framework, global footprint map, risk watchlist, and weekly narrative layer to surface which upstream constraints matter most for AI deployment planning.

AI Infrastructure Research Portfolio — Curated HTML portfolio of my AI infrastructure research and visual storytelling work.

Postmates LA Retention Analysis — User retention & frequency strategy - Analyzed retention patterns across user segments, merchant mix, delivery experience, and promo dependency for Postmates' LA market. Found that local and trend-driven merchants drive 90–96% repeat user share vs. 85–88% for generic merchants; modeled a one-time user conversion opportunity worth ~$4K incremental GMV and built a strategic playbook for the LA market.

Nazava Shopee Sales Optimization — E-commerce revenue strategy Analyzed 16 months of Shopee sales data for a water filter company. Identified that 73% of revenue comes from just 2 SKUs, built a promotion timing model that projected 35% revenue increase, and designed a bundle pricing strategy from cross-sell patterns.

InVehicle Coupon Strategy — Coupon targeting with ML Built an XGBoost model (AUC 0.832) on 12,684 driving scenarios to predict coupon acceptance. Segmented drivers into 5 behavioral personas — targeting the right ones cuts coupon waste by 41% while maintaining 89% reach of likely acceptors.

Content Creation Agent — Research-backed video script pipeline Web search → fact extraction → cross-domain analysis → script generation. Turns 40+ articles into a 10-min YouTube script with cited sources in 4 minutes.

Yapyap-AI — Voice conversation practice Gamified sandbox for practicing conversation skills with AI voice feedback.

Anthology-AI — Time-aware news mapping Tracks how topics evolve across years by mapping scattered news into structured timelines.

I make data-backed analysis videos on AI infrastructure — chips, data centers, energy, and market dynamics.

Latest:

• Understand Jensen Huang's GTC Keynote = Understand the Future of AI — Breaks down the 2026 GTC keynote: Vera Rubin, Kyber architecture, OpenClaw, and what it signals about where NVIDIA is heading | Read the article
• The 3 Biggest Bottlenecks for AI | Why AGI Is Further Than You Think — Power, cooling, and chip supply constraints that determine the actual pace of AI scaling | Read the article

The Content Creation Agent above is built to support this workflow: web research → fact extraction → script generation, all with human review at every step.

In the age of AI, the bottleneck isn't coding—it's clarity of thought and speed of iteration.

I believe we're in the early stages of a fundamental shift in how intelligence works. Instead of predicting what happens next, I'm building things that help me understand it better—tools for content creation, research, and analysis at both the application and infrastructure layers.

Every project starts with identifying where human judgment adds value, then uses AI to handle everything else.

LinkedIn • Santa Clara, CA