About me
I'm a product and data leader with 10+ years across SaaS, startups and impact investing.
At NetDocuments I own the analytics platform behind the product roadmap, and I build AI agents on Snowflake and dbt. Before that I co-founded a Berkeley SkyDeck-backed AI startup for YouTube creators, led product, and grew it to 2,000 users and $140K in revenue.
How I work on product
Customer discovery
At my startup I ran 80+ interviews with YouTube creators to set the roadmap. The top two pains were competitor analysis and video ideas. We built video ideas first because it was the more achievable of the two.
Launches and experiments
I led three AI tools (ideas, keywords, thumbnails) from discovery to launch, reaching 2,000 active users in 14 months. I designed the A/B test for the thumbnail tool myself: randomized by impression, 60+ creators, 8,000+ videos. Click-through rose 30%.
Iterating from usage data
Usage logs and feedback showed 20% of our AI suggestions were repeats. I rewrote the prompts to use each channel’s own data, and ratings went from 3.2 to 4.3 out of 5.
Metrics and requirements
At NetDocuments I worked with Product and Customer Success to define adoption as depth plus breadth, and wrote the user stories and business cases for AI hackathon builds that shipped.
What I build myself
Data modeling
Snowflake and dbt. Dimensional models, a semantic layer, and tests for uniqueness, reconciliation and lineage, so one model can serve dashboards, notebooks and AI agents.
AI agents in production
I built and shipped a planning app with Next.js, FastAPI and the Claude Agent SDK. Agents answer questions over governed semantic models, with evals behind them. When the data cannot answer, they flag the gap instead of guessing.
Automation
Python agents on the Power BI REST API and Slack that catch data-refresh failures, find the cause and recommend a fix.
Machine learning
Fine-tuned a Hugging Face image model on 2M+ YouTube thumbnails to score new thumbnails and suggest fixes, served through Python and Django.
StackSQL, Python, dbt, Snowflake, Power BI, Hex, Pendo, Next.js, FastAPI, Claude Code, Claude Agent SDK
CertificationsSnowPro Core (2026), Google Advanced Data Analytics. dbt Analytics Engineering exam in December 2026.
Background
- 2025 to now
Product analytics and AI agents
NetDocuments, Utah
- 2023 to 2025
Co-founder and Head of Product
Upview AI, an AI analytics startup for YouTube creators. Berkeley SkyDeck and CMU VentureBridge.
- 2019 to 2023
Chief of Staff to the CEO
A family office in impact investing. Owned the $12M annual funding process for a $180M portfolio.
- 2017 to 2019
Master of Development Practice
UC Berkeley
- 2015 to 2017
Head of Strategy, founding team
Skillsphere Education, Mumbai. Designed programs for grades 6 to 12 and trained 120 school faculty.
- 2013 to 2015
Tech Analyst, Advisory Services
EY, Mumbai
Outside work
I teach. I started in classrooms and now make videos about using AI on real business data. I live in Utah, play badminton, and am learning photography on a Sony A7 II.