Aug 2025 – Present · 8 mos · Mountain View, CA

• 🚀Everything AI-related end to end for Core Product Team.
• Gatekeeper for prompts and eval strategy across work streams. Codeowner over some questionable backend code and every system prompt.
• Recruiting and calibrating for our first early career talent cohort! DM me.
• Contributions (surface wise):
• Growth, Connectors, Agents (almost every core behavior of Copilot Tasks)

Aug 2025 – Present · 8 mos · Mountain View, California, United States · On-site

• Needed this to add work verification for Microsoft

Jan 2025 – Jan 2026 · 1 yr · San Francisco Bay Area

• Figma-to-codebase generation with pixel-perfect Tailwind output. Generated millions of impressions, added thousands of businesses or users to the waitlist and convinced several angel investors. Created an npm package, canvas editor, repo importer, language analyzer, partial html/css unwinding, AI generated storybooks, and more. Shuttered operations after my co-founder decided to pursue other things. We chatted with a16z if that means anything 🤷‍♂️

Jul 2024 – Aug 2025 · 1 yr 1 mo · Vancouver, BC · Remote

• 💵 Led the receipt OCR discrepancy project (operating on 86% of Instacart orders) from under 50% to 90% recall by integrating LLMs into traditional data mining ML. Owned the entire microservice, data pipelines, async jobs, algorithm development, datadog alerts and some data analyses. This problem was considered impossible with high fidelity for years and saves $2-3M a month. Planned improvements to expand the impact of this system to $10M+ a month given 98% precision - creating more algorithms and pipelines for product pricing, fraud and order quality work streams.
• Executed one of the first LLM integrations in fulfillment, optimized pyspark ML pipelines to be 20x faster, decreased system latency from 30 secs to 3 secs, built human eval system for receipt extraction and item matching.
• Promoted before the fastest time possible, gaining an exception due to ranking within the top 5% of performers within my first 6 months (took 3 months for the promotion cycle to finish).
• Also helped across card payments (Stripe, Marqeta), transaction rejections and tax suppression with millisecond and 99.99% reliability SLAs. Bonus: implemented company wide log sampling for python microservices.

Jan 2024 – Jan 2024 · 0 mo

• 🎥⏺️ AI agents that operate computers. We were a bit too ahead of our time before models could work effectively in this space and sales didn't quite work out. Had some angels and YC wanted to chat at least 🤷‍♂️. Full credits to Charles for being the mastermind behind the idea and v0.

Jan 2023 – Jan 2024 · 1 yr · Vancouver, BC

• 🧬 Deep learning, bayesian methods, HPCs, computer vision, and some full stack development for single cell sequencing at Canada's Michael Smith Genome Sciences Center (founded by the Nobel Prize Winner).
• Yes, it was founded by the Nobel Prize Winner before he passed, read the backgrounder: https://archive.news.gov.bc.ca/releases/news_releases_2020-2024/2023HLTH0035-000911.htm

Jan 2020 – Jan 2024 · 4 yrs · Vancouver, British Columbia, Canada

• 🧑‍🎓 Maintainer for MOOC Micromaster programs for Software Development in Racket (sub languages) and TypeScript having over registered 188,700 students. Featured on OSSU Computer Science GitHub repository, which has 170k+ stars.
• Mainly helped introduction to programming (CPSC 110) solving problems such as graphs and search for ~700 students per term. Actively led labs and made 3k+ contributions to question threads on programming problems as complex as generative recursion, search, and graphs.
• Also helped with intermediate algorithm design and analysis (CPSC 320) for graphs, greedy, DP, and NP problems for ~ 200 students.
• Made the main autograding systems for statistical modelling in data science (STAT 301).
• Served for ~10 academic terms.