Projects, case studies, and a few too many slides.
A collection of things I've built, shipped, and figured out along the way.
Research, strategy, and a lot of late nights. Some polished, some messy—it's all part of the journey.
PS: Some of these have too much text on the slides. I know. I was learning.
GPU Cost Intelligence
Cornell University (Cornell Tech) Product Fellowship
This project went through multiple pivots before landing on GPU pricing. Started with interconnection queues in renewable energy—urgent problem, but oversaturated and too policy-driven. Pivoted to generative design for data centers, exploring how AI could optimize physical infrastructure for energy efficiency. Again, high barriers and market saturation. The final pivot came from a real insight: as demand for generative AI explodes, GPU infrastructure has become the new bottleneck—and the pricing landscape across cloud providers is fragmented, volatile, and largely opaque.
Led product research and built a real-time GPU pricing intelligence pipeline to help AI startups make smarter infrastructure decisions. Discovered 40-60% cost variance across cloud providers—pricing chaos that most teams don't even know exists.
Tech: Python, Google Cloud Billing API, Web scraping, Selenium
Gender Bias in ML Credit Scoring
NYU Research Project
Mixed-methods research investigating how machine learning credit-scoring models perpetuate gender bias. Evaluated fairness-aware techniques like adversarial debiasing and subgroup threshold optimization, combining quantitative analysis with qualitative interviews with financial professionals.
(This project involves the usage of AI for research and writing)
UN AI Media Analyzer
NYU Capstone for United Nations
Built a multilingual AI tool for the United Nations to detect hate speech, misinformation, and toxicity across audio, video, and text. The tool supports 6 official UN languages and was developed in collaboration with the United Nations International Computing Centre (UNICC) to support ethical media monitoring at a global scale.
Tech: ChatGPT, Claude, Whisper ASR, XLM-RoBERTa, BART, Flask, OpenCV, Gemini API
InsightGen
GenAI Competitive Intelligence Tool | A time before ChatGPT could perform real time search
An AI-powered productivity tool that transforms fragmented, scattered data into actionable insights for marketing and PR professionals. Built with real-time dashboards, keyword tracking, and a multi-modal chatbot—designed to help teams stop drowning in data and start making decisions.
Tech: Vertex AI, Gemini, Google Cloud
HelloFresh Product Innovation
Product Strategy Project
Developed a product strategy to tackle HelloFresh's churn problem through AI-driven meal personalization and sustainable packaging. Proposed an "Express Meals" line for time-crunched customers and LLM-powered recipe simplification to make cooking less intimidating.
Focus: Express Meals, eco-friendly packaging, LLM-powered recipe generation
Abacus Bank Revival
Business Strategy Case Study
Created a strategic turnaround plan for Abacus Bank—the only bank headquartered in NYC's Chinatown, serving Chinese immigrants since 1984. Proposed digital transformation, student marketing at local universities, and new financial products tailored to immigration needs.