Zackary Santana

About Zackary Santana

I am a full-stack developer that loves building for users. My portfolio is at https://attach.fly.dev/My preferred languages are Golang, TypeScript, or Java. I have worked with a microservice architecture with 30+ microservices, a monolithic architecture with 500k lines, a serverless architecture with AWS Lambda functions (SST), a traditional frontend team and backend team that use GraphQL to communicate, and a distributed systems team for internal CI/CD and testing infrastructure.I enjoy learning about different ways to solve the same problem and which way might lead to the best DX, performance, or velocity!

Experience

1. Software Engineer Iii

   MongoDB

   Feb 2025 — Present · Miami, FL, US

   Implemented test regression bisection logic, reducing tests run per day by 30k (15k compute hours), which reduced costs by $43k/month (~$525k/year) while accelerating developer velocity. Parallelized git cloning across test infrastructure, reducing setup time from 2m → 30s for a targeted 200k+ daily test runs, saving compute hours/day (~$175k/year) and significantly accelerating baseline feedback speed. Replaced static, shared AWS credentials with on-demand temporary credentials issued via STS, enforcing least-privilege access, full auditability, and eliminating persistent secret exposure across 1 million tests daily. Replaced reliance on shared instance profiles for AWS access by introducing application-server minted temporary credentials, enabling end users across 40 services to securely AssumeRole instead of using static credentials. Replaced long-lived GitHub OAuth tokens with short-lived GitHub App installation tokens, enforcing least-privilege access, automatic expiration, and eliminating persistent credential exposure across 1 million tests daily and internal jobs daily. Built a Visual Studio Code extension that allowed engineers to create and modify test runs, examine results, and inspect test run changes directly inside the editor. Dogfooded the latest MongoDB Go driver, identifying and reporting two critical regressions before they reached end users. Collaborated across teams to dogfood upcoming MongoDB Atlas server release candidates, validating functionality, and stability in our staging and production environments. Expanded distributed tracing coverage in Go applications, enabling faster outage debugging and early detection of performance regressions (e.g, missing database indexes). Augmented observability with disk usage and network metrics, allowing the team to identify over-provisioned resources, such as tests using unnecessarily large instance types or excess EBS volumes, and optimize infrastructure efficiency.

Education

• Florida International University

  Bachelor of Science - BS, Computer Science

• Terra Environmental Research Institute

  High School Diploma, Environmental Science

  2016 — 2020