Jimmie has deep expertise in developing back-end services in Go and managing and automating infrastructure with IaC. He specializes in Go, Google Cloud, Kubernetes, and Terraform. He has used these for over five years to build back-end services and CI/CD pipelines and improve automation. In one of Jimmie's most exciting projects, he developed cloud infrastructure for a central eCommerce platform that serves students at more than 900 college campuses.
Refactored testing and test data for readability and usability. Created a helper library to define data required for a test declaratively, add it to the database, and clean it up afterward.
Designed and implemented a generation of zip files for user download.
Redesigned and rolled out access control changes to work with organizational feature requirements and the addition of the organization's related APIs.
Diagnosed issues causing bloated bundle size and reduced react app bundle size by 75%.
Improved lighthouse score by 50+ for desktop load times through caching and front-end optimization.
Organized the GitLab CI/CD for deploying microservices and infrastructure changes.
Created Terraform scripts to manage infrastructure as code (IaC).
Designed, set up, and managed a per-branch UI preview, allowing QA to go through a branch with front-end changes before merging to a staging environment.
Provided actionable insights about performance issues with the browser SPA and did some development to implement fixes in this area.
Managed and set up monitoring and alerts to quickly flag issues affecting customers.
Implemented OCM's caching strategy. At its core, requests to the server never include user data, which works well for this project and allows CDN to store all results returned from the back-end until code or data changes.
Observed and fixed several memory leaks in the react application.
Developed a website designer that exports pixel-perfect React code every time. Drag and drop React components were implemented using the MUI design system. It's a design tool that works seamlessly with your development processes and eliminates integration work. I'm the only person working on this product, so I wear all the hats.
The research focused on generative design without an input dataset. Attempt to generate interesting floor plans, schematic layouts, and efficient box packing.Built graph neural networks, fully differentiable loss functions, investigating non-gradient descent-based deep learning.
A technical blog on various topics, including terraform, machine learning, Go, and Kubernetes.Some highlights include:• Setting up Kubernetes zero-downtime deployments with testing.• Zero loss on the fly expansion of linear layers deep learning.• Additional posts around deep learning research I've been working on.
