- Led re-architecture of a multi-region QA platform (US, CA, EU, AU) with EU data residency and per-region isolation on a unified codebase, and built a CDC pipeline streaming PostgreSQL data into Snowflake for business analysts
- Relocated the EU Cloud Run service from the US to the same region as its Frankfurt database, cutting service median latency 96% (~500ms → ~19ms) by eliminating transatlantic round-trips
- Engineered a kitbag bundling algorithm as part of the P500 initiative that cut cart runs by 6% and developed auto-cart picking functionality, directly increasing warehouse throughput and picker efficiency
- Built and owned Redzone Product Master, automating synchronization of source-of-truth data into the Redzone database and eliminating manual data entry
- Delivered an automated email notification system for the QA App to surface overdue and updated tasks, strengthening audit accountability and reducing manual follow-up overhead (Python, FastAPI, PostgreSQL, GCP, SendGrid)
- Improved developer experience and system observability across multiple services by integrating API Gateways, tracing, structured logging, dashboards and automated linting standards into the team's stack
MichaelGee.
Software engineer with 5+ years shipping production systems at HelloFresh and ZoomInfo - multi-region platforms, bundling algorithms, and microservices that cut latency and defects. I pair distributed-systems and database fundamentals with product taste, focused on agentic workflows.
A bit about me
I'm a software engineer with 5+ years of experience, currently at HelloFresh, where I work on backend systems and platform architecture. I led the updated architecture of a multi-region QA platform serving the US, Canada, EU, and Australia - the kind of work where the hard part isn't the code, it's the product and organizational calls that keep one codebase sane across four regions.
I like building things that move a number. At HelloFresh I shipped a bundling algorithm that cut warehouse cart runs by 6%, a FastAPI service that eliminated heaps of manual data entry into Redzone, and auto-picking functionality that raised kitbag cart-run throughput. Before that, three years at ZoomInfo on their sales-engagement platform Engage - cutting p95 latency by 43%, building backend microservices for the AI Emailer, and full-stack work across the frontend. That's where I got comfortable owning features end to end, across the stack.
What I care about most now is the product and the people building it. I'm at my best turning fuzzy business goals into systems that ship, making the calls on architecture and scope, and bringing a team with me. That's the kind of leadership I care about - and I hold the bar on code quality just as high while doing it. Product taste and clear communication amplify strong engineering; they don't replace it.
Where I've worked
- Built Engage AI Emailer backend microservices, including throttler and REST API, to increase engagement capabilities for users (GCP, Node.js, NestJS, Typescript, MongoDB, Jest, Prisma, Kafka, Docker)
- Owned multiple hotfixes such as implementing AWS FIFO queue to prevent duplicate contact tasks within Salesflows leading to 40% reduction of high priority defects for Engage platform
- Completed engineering excellence tasks and fixed software issues found during on-call rotation and production deployment leading to 23% increase in MAUs for Engage
- Reduced p95 latency by 43% by detecting bottlenecks, implementing new asynchronous data retrieval endpoints, and writing efficient SQL queries (.NET Core, C#, Datadog, SQL)
- Built infrastructure to handle persistence and manipulation of form data for the Salesflow Wizard Microfrontend
- Built and demoed PoC of Apache Airflow by implementing a REST API that uses JWT authenticated endpoints to handle user data, designing Airflow workflow, and developing UI to illustrate data manipulation (Node.js, Sqlite)
- Built a single page application that uses company internal API to decrease time spent by business in tracking location data (React, Storybook)
What I'm building
- Developing a full stack app with an Anthropic-powered conversational assistant that generates multiple coordinated schedule changes in one request (Next.js, TypeScript, Python, FastAPI, Supabase, Anthropic API, Google Calendar API)
What I work with
Off the clock
Poker
Reading the table, managing variance, knowing when to fold. Math and psychology both scratch an itch.
Chess
Openings, tactics, and the long game - where patience meets pattern-recognition.
Lifting
Progressive overload and showing up. Discipline under the bar that carries into everything.
Cooking
Weeknight experiments to from-scratch meals. Good food is a system worth iterating on.
Gaming
Strategy, competition, and worlds worth getting lost in after a long day of building.