Tsinu A.
Gezahgn

Build and maintain APIs for the Configuration Blueprint platform, which controls how software packages are defined, versioned, and deployed across JPMorgan's private cloud. Led automation work that cut deployment time by 30 minutes per package and shipped tooling that moves teams off deprecated configurations at scale. Also exploring ML and GenAI use cases, including semantic search with AWS Bedrock, to reduce misconfiguration rates across cloud environments.

Led a team of five engineers to break up a monolithic job advertisement system into event-driven serverless microservices across 15+ AWS services. Automated infrastructure with CDK, implemented CI/CD pipelines that tightened the release cycle, and hardened backend security with JWT authentication and TLS certificate management.

Built and maintained a Java Security SDK with mTLS, OAuth 2.0, and JWT filters to standardize secure app-to-app authentication across Apple's Ad Platform microservices. This reduced security protocol implementation time by 40% and cut security incidents by 50% per internal audits. Also migrated services to Spring WebFlux for non-blocking throughput and worked through OWASP vulnerabilities with continuous SonarQube monitoring.

Rebuilt the company's core analytics product from the ground up, cutting data processing latency from 24 hours to three minutes with a PostgreSQL-backed event-driven read model. Migrated job servers from dedicated EC2 to serverless (ECS Fargate, Step Functions, Docker), which cut infrastructure costs by 40% and improved operational efficiency by 30%. Also built an Integration-as-a-Service system that let clients push raw CSV emissions data directly into the platform, removing the need for manual conversion and reducing integration effort by 40%.

Designed and implemented a Java 8 parallel dependency library for multi-core architectures, working through concurrency models and task-scheduling strategies for high-throughput workloads. The work gave me a grounding in parallel computation that still shapes how I think about performance-critical systems.

Built a vulnerability record system to track and report security issues across government infrastructure, did penetration testing on critical systems, and wrote a custom fuzzer to detect buffer overflow vulnerabilities. That hands-on offensive security work gave me a lasting understanding of how systems break under adversarial conditions.