Senior Software Engineer at Google
Software engineer,
proud owner of two cats.
I’m Eric, a Senior Software Engineer at Google, building the backend, data and applied-AI systems behind the platform 30,000+ Google Ads sellers use every day. I started out in life sciences at the University of Toronto, with a BSc in Health & Disease and time as an Immunology Research Student at UHN, before a Master’s in Biostatistics focused on AI.
01The journey
From immunology research, to biostatistics, to building AI systems at Google.
Sep 2022 – Present
Senior Software Engineer · Sales CRM · Mountain View, CA
Backend tech lead for the Sales CRM Portfolio team of eight engineers: setting backend technical direction, scoping and distributing work, mentoring engineers and driving OKR planning. Sales CRM is the platform 30,000+ Google Ads sellers use to manage $200B+ in business.
Sep 2021 – Sep 2022
Practicum Research Student · Dr. Pingzhao Hu
Unsupervised clustering of single-cell RNA-seq data with graph autoencoders. First author on scGMM-VGAE, a Gaussian-mixture-model variational graph autoencoder for clustering scRNA-seq data.
Sep 2019 – Aug 2021
Research & Teaching Assistant · Research Opportunity Program
Simulated how immunization and social distancing shape the spread of COVID-19 using COBWEB, an agent-based modelling tool. The work was featured by the Faculty of Arts & Science. Later supported undergraduate students using COBWEB in their own research and ran help sessions.
Read the feature ↗Sep 2019 – Apr 2021
Immunology Research Student · Dr. Bryan Coburn
Studied how IgA relates to HIV susceptibility, then the bacteria linked to chronic lung allograft dysfunction (CLAD) and how to detect bacteria-specific IgA. Ran ELISA, DNA extraction and qPCR experiments, and analysed and visualised the results in R.
2021 – 2022
University of Toronto · Emphasis in Artificial Intelligence & Data Science
2017 – 2021
University of Toronto · Minors in Statistics and Immunology
02The work
Most of this work runs inside Google, so there’s nothing to click through. Each one covers what I owned and what it changed.
Google · Agentic AI · Generative UI
A 0-to-1 agentic workspace that turns live sales data into interactive canvases. I led backend generation for Google’s A2UI framework, building agents that emit declarative UI with data-binding and streaming contracts, and co-built the orchestration of 15+ agents and 25+ skills on top of 25+ partner MCP tools.
Google · LLM systems · Backend
An LLM system that reads 1.9M+ advertiser websites and extracts business descriptions, social links, conversion actions and product catalogs. I owned the backend: a Java gRPC service for on-demand generation, a refresh server for new and stale sites, and pre-processing that cut HTML input tokens by 81%.
Google · Machine learning · Ranking
An ML model that ranks each seller’s accounts to surface the highest-value opportunities. I refined the ranking by blending account signals with business rules, and used ML interpretability to explain every ranking to the sellers acting on it.
Won GM’s “Putting EVerybody in an EV” prize at UofTHacks VIII. A web app for renting out home EV chargers and finding one nearby. Express, MongoDB, Google Maps API.
Finalist at Hack the North 2020++. Counts people entering and leaving a store from camera footage using YOLOv3 and OpenCV, with an Express and MongoDB backend on Azure.
Decision-tree and TensorFlow models trained on four decades of team statistics. Its close runner-up, the Milwaukee Bucks, went on to win the 2021 title.
A Discord bot for looking up University of Toronto courses and exam schedules, built on Node.js and the Nikel API. It served more than 4,300 Discord users.
A multi-purpose Discord bot covering music, chat-based ranking, daily Ontario COVID-19 updates, and League of Legends and stock lookups through the Riot and IEX APIs.
03The toolkit
Backend
Data
Applied AI
Modelling
Code