Dive log LC-25Lucas Chen · descentRésumé

Dive log LC-25 · field guide to a software engineer

Lucas Chen

Software Engineer at Zynga (AdTech) · Toronto

I'm a software engineer on Zynga's AdTech team in Toronto. I designed and led our internal tooling for ad-serving configurations, an MCP (Model Context Protocol) server and tools that let AI assistants work with ad-serving configuration, and a project that helps our games identify high-value ads. I also worked on an AI LiveOps agent that cut a 2–3 hour workflow to about 20 minutes.

Read the dive logGitHubLinkedIn

Porthole · surface · 0 m
Diver
Lucas ChenSoftware Engineer
Dive site
Zynga · AdTechSince May 2025 · 1 yr 4 mo down
Base
Toronto, Canada43.65° N, 79.38° W
Certification
McMaster UniversityB.Eng., Software Engineering · 2025
Log so far
6 dives · 10 specimens5 featured
Begin the descent

Sunlit zone · 0–200 mExperience

The dive log

Every role since Jun 2021, newest first. The dive still in progress is at the top.

  1. Dive no.
    06
    Date
    May 2025 – Present
    Site
    ZyngaToronto, ON
    Bottom time
    1 yr 4 mo

    Software Engineer

    Dive in progress
    • Designed and led internal tooling for ad-serving configurations: an MCP (Model Context Protocol) server and tools that let AI assistants work with ad-serving configuration.
    • Worked on an AI LiveOps agent that cut a workflow from 2–3 hours to about 20 minutes.
    • Designed and led a project that helps games identify high-value ads.
    Profile · bottom time to scale
  2. Dive no.
    05
    Date
    May 2023 – Jun 2024
    Site
    Hexagon Manufacturing Intelligence
    Bottom time
    1 yr 2 mo

    Full Stack Engineering Intern

    Internship
    • Developed 15+ UI and 10+ API production features for the Nexus Platform, using React and Express with TypeScript.
    • Reduced load times for user list pages by 97% by optimizing API calls and refining Express APIs.
    • Cut test automation start-up time by up to 98% by parallelizing the Playwright test suite setup.
    • Reduced API response time by up to 82% with an Azure Cosmos DB proxy that moved Express APIs to C# .NET database pagination.

    GearTypeScriptReactExpressC#.NETAzure Cosmos DBPlaywright

    Profile · bottom time to scale
  3. Dive no.
    04
    Date
    Sep 2022 – Apr 2023
    Site
    McMaster UniversityHamilton, ON
    Bottom time
    8 mo

    Teaching Assistant

    Teaching
    • 1P13 Integrated Engineering Project (Sep–Dec 2022).
    • 2GA3 Computer Architecture (Jan–Apr 2023).
    Profile · bottom time to scale
  4. Dive no.
    03
    Date
    May 2022 – Aug 2022
    Site
    Exchange Solutions
    Bottom time
    4 mo

    Software Engineering Intern

    Internship
    • Made e-commerce client onboarding over 10x faster by building AWS Lambda components.
    • Led an augmented reality proof of concept for a national loyalty program, built with 8th Wall WebAR and JavaScript.
    • Raised a loyalty platform’s user engagement by 4% by integrating it with blockchain assets, using Next.js, Ethereum and the Alchemy API.
    • Built a real-time data visualization dashboard (D3.js, Snowflake), improving early error catching and data comparisons by over 10%.

    GearAWS LambdaJavaScriptNext.jsD3.jsSnowflake8th Wall WebARShopify API

    Profile · bottom time to scale
  5. Dive no.
    02
    Date
    Oct 2021 – 2022
    Site
    McMaster Artificial Intelligence SocietyHamilton, ON
    Bottom time
    n/a

    AI Research Member

    Research
    • Worked with a team of four developers on training a humanoid to walk and navigate obstacles in a ragdoll physics simulation.
    • Used C# and Unity to train and simulate the humanoid.

    GearC#Unity

    Profile · bottom time to scale
  6. Dive no.
    01
    Date
    Jun 2021 – Sep 2021
    Site
    InsideDesk
    Bottom time
    4 mo

    Software Engineering Intern

    Internship
    • Found 50+ bugs through QA testing, validated bug fixes, and was the link between clients and the development team.
    • Cut manual tasks by over 30% with custom JavaScript Slack bots that automate the customer onboarding notifications.
    • Wrote automated end-to-end UI tests with Selenium and Python, cutting manual QA testing by 40%.

    GearJavaScriptPythonSeleniumSlack

    Profile · bottom time to scale

Twilight zone · 200–1,000 mImpact

Discoveries

Found at Zynga AdTech, where the light gives out: 3 discoveries, 2 of them designed and led by Lucas.

  1. Field no. D-01Twilight zone

    Designed & led

    Ad-config MCP

    I designed and led internal tooling for ad-serving configurations at Zynga: a Model Context Protocol (MCP) server and tools that let AI assistants work with ad-serving configuration.

    • AI
    • MCP
    • AdTech
    • Leadership
  2. Field no. D-02Twilight zone

    Worked on it

    LiveOps agent

    I worked on an AI LiveOps agent that cut a workflow from 2–3 hours to about 20 minutes.

    Light meter · Time to complete one LiveOps workflow
    Before2–3 h
    After~20 min
    • AI agents
    • LiveOps
    • Automation
  3. Field no. D-03Twilight zone

    Designed & led

    High-value ads

    I designed and led a project that helps games identify high-value ads.

    • AdTech
    • Leadership

Midnight zone · 1,000–4,000 mProjects

Specimens collected

10 projects catalogued since 2022; 5 on display below, the rest in the register.

  1. Specimen SP-01 · Tool

    Agent Platform

    A control plane for safely delegating software work to autonomous coding agents.

    Collected
    2026
    Habitat
    Aug 2026 – present
    Collector
    Solo: spec, design and CLI (in progress)
    Feeds on
    TypeScriptNode.jsJSON Schema (Ajv)CommanderYAML
    • Wrote the v3 spec: missions, workflow templates, task and role contracts, review gates, model routing, memory and fleet security.
    • Built the “agent” control-plane CLI in TypeScript: it scaffolds a target repo, validates artifacts against JSON Schemas and runs the claim → start → submit → gate loop.
    • Workers never certify their own work: a task passes only with an evidence bundle, independent verification, and spec, quality and architecture review.
    • Routes each task by risk to a capability tier (frontier, strong, mid, cheap), mapped to models by policy instead of hard-coded vendors.
  2. Specimen SP-02 · Product

    TennisBooker

    Books tennis and pickleball courts the moment booking opens, every morning.

    Collected
    2026
    Habitat
    2023 – present
    Collector
    Solo, full-stack (2023 version with help from Douglas Chen)
    Feeds on
    TypeScriptPlaywrightReactExpressPostgreSQLAuth0Docker
    • The bot logs in at 7:25, builds every reservation, then fires at exactly 7:30 a.m. ET for courts 7 days out, falls back to backup courts and texts results via Twilio.
    • TypeScript monorepo (Turborepo): Express + Prisma/PostgreSQL API with Auth0 roles and machine-to-machine tokens for the bot, and a React, MUI and TanStack Query web app.
    • Runs in Docker on a Raspberry Pi: CI builds multi-arch images to GHCR, deploys go staging → smoke test → promote with one-command rollback, and Vitest and Playwright suites cover the code.
    • Began in 2023 as “Alfred”: a Playwright bot on a cron-scheduled GitLab pipeline, with a React and Express app, Firebase and multi-user Auth0 login, used by 2 clients. In Aug 2026 the bot moved to the booking platform’s new GraphQL API.
  3. Specimen SP-03 · Product

    PlatefulAI

    Mobile meal planner: pick meals, get a grocery list, track calories and protein.

    Collected
    2024
    Habitat
    Nov – Dec 2024
    Collector
    Team of 5: app shell, Auth0 login, meal-planning flow, deploy
    Feeds on
    ReactViteMUIAuth0Chart.jsVercel
    • Set up the React + Vite app with Auth0 login, the start and home screens and bottom-tab navigation, and deployed it on Vercel.
    • Built the meal-planning flow: favourites, filtering out meals that clash with saved allergies and diets, a meal-details view, and meal and serving selection.
    • Teammates built the grocery list and meal data (Mohammad Mohsin Khan), the calorie and protein tracker (Wajdan Faheem), and profile and target settings (Jasmine Sun-Hu).
  4. Specimen SP-07 · Product

    CarpoolX

    Android app that matches riders headed the same way so they can share a taxi.

    Collected
    2023
    Habitat
    Mar – Apr 2023
    Collector
    Team of 4: repo lead, most of the Flask backend and auth, core screens
    Feeds on
    React NativePythonFlaskFirebaseGoogle Maps Platform
    • Led the repo: scaffold, navigation and a feature-branch workflow. Merged all 38 pull requests that landed and wrote 119 of 156 commits.
    • Wrote most of the MVC Flask backend: multi-user Firebase Auth sign-in, tokens kept in the Android Keystore, token-checked endpoints, and the controller and model layers.
    • Built login and registration, the profile, taxi verification by door code, taxi details and passenger reports.
    • Teammates built ride requests and matching with Google Maps routes and fare estimates (Henushan Bala), end-of-ride ratings (Rafeed Iqbal) and private group rides (Allison Cook).
  5. Specimen SP-10 · Hackathon2nd place, “To Work or Not To Workout” sponsor challenge (AI Endurance), DeltaHacks 8

    Rest Day

    A workout planner that schedules rest days, not just workouts. Built at DeltaHacks 8.

    Collected
    2022
    Habitat
    Jan 2022
    Collector
    Team of 2: Fitbit API integration, landing page, deploy
    Feeds on
    ReactFitbit Web APIstyled-componentsNode.jsFigma
    • Wrote the Fitbit Web API client, which pulls 7-day sleep and activity logs and heart-rate data (a standalone Node script, not yet wired into the UI).
    • Built the navigation bar and landing page with styled-components, plus the routes for the other pages.
    • Merged the feature branches, deployed the React app to GitHub Pages, and wrote the README and Devpost story.
    • Willie Pai built the weekly schedule view and the exercise GIFs.

Register of other specimens5 more in storage

  1. SP-04StudySpotFind a place to study at McMaster: search, filters, ratings, favourites and a map.Product · Nov 2024Live demo ↗
  2. SP-05UnPasswordA password manager built in one day for the MEC 2024 programming challenge.Hackathon · Nov 2024GitHub ↗
  3. SP-06API ResumeMy resume as a REST API: Postman-only, rate-limited, with a joke for anyone who POSTs.Experiment · Oct 2023Field notes →
  4. SP-08Ride HomeMix subway sounds, chatter and lo-fi music into your own TTC ride-home ambience.Experiment · Jun 2022Live ↗
  5. SP-09Summary.ioUpload a lecture video and get an AI transcript to study from. Built at MacHacks 2.Hackathon · Jan 2022Devpost ↗

The abyss · 4,000–6,000 mSkills

Equipment check

The pre-dive checklist: every kit that goes down with the diver, and how much is in it.

  1. Kit 01

    Languages

    8 tools checked

    • TypeScript
    • JavaScript
    • Python
    • Java
    • C#
    • SQL
    • HTML/CSS
    • Bash
  2. Kit 02

    Frontend

    6 tools checked

    • React
    • React Native
    • Next.js
    • Vite
    • MUI
    • TanStack Query
  3. Kit 03

    Backend & data

    11 tools checked

    • Node.js
    • Express
    • .NET
    • Flask
    • Prisma
    • PostgreSQL
    • MongoDB
    • Azure Cosmos DB
    • Firebase
    • DynamoDB
    • Swagger
  4. Kit 04

    AI & agents

    4 tools checked

    • Model Context Protocol (MCP)
    • Claude Agent SDK
    • Claude Code
    • Cursor
  5. Kit 05

    Testing & automation

    4 tools checked

    • Playwright
    • Vitest
    • Jest
    • Selenium
  6. Kit 06

    Infra & tooling

    8 tools checked

    • Docker
    • Git
    • GitHub Actions
    • GitLab CI
    • Vercel
    • AWS (Lambda, S3)
    • Auth0
    • Azure DevOps

Certification cardCertified

B.Eng., Software Engineering (Co-op)McMaster University · 2020–2025Dean's Honour List (2020–2023) · McMaster Honour Award (2020–2023) · Courses: System Design, Data Structures and Algorithms, OOP, Software Design, Databases

Diver: Lucas Chen

All equipment checked6 of 6 kits · 41 tools ready for the descent

Sub dock · 6,000–6,500 mContact

Resurface

The dive is done and the sub is docked in Toronto. Take an ascent line: pick a buoy and Lucas will pick up.

Why this portfolio?

The Deep sea dive edition: the same facts as every other version, in a different shape.

How each section of this page is drawn
IntroPorthole at the surface
ExperienceDive log entries
ImpactDiscoveries in the twilight
ProjectsCatalogued specimens
SkillsPre-dive equipment check
ContactResurface at the dock

Midnight palette · Survey type · Kinetic motion

Remix
How this site works →