Java and Spring services that sit in the middle of real work.
Backend systems
Day to day I build and maintain Spring services responsible for document-processing workflows, APIs, integrations and production workloads. That usually means a boring-looking service with an un-boring job: accept work from other systems, keep a consistent record of it, and hand it on without losing the plot. I care about the shape of the API, the data model underneath it, and what happens when a downstream dependency is slow, partial or wrong.
Files that are too big to pretend they are JSON.
Large-file architecture
Some of the more interesting problems have been around large files: object storage, metadata that has to stay searchable, streaming rather than loading, and integrations that must remain backwards-compatible while the storage model changes. The constraint is rarely 'can we upload this'. It is 'can another service, written years ago, still find it, and can we change the path without a ceremony'.
Most failures do not live in one process.
Production debugging
When something breaks in production it is often not a stack trace in a single service. It is a story that crosses HTTP, a database, a message, some cloud infrastructure, and an assumption two teams no longer share. I like that work. The method is slow on purpose: reproduce what you can, narrow the boundary, prove where the data last looked right, and only then change code. Guessing feels faster. It usually is not.
Models are useful when they have a job and a place to put the answer.
AI integrations
I have integrated AI-powered document classification and information extraction into existing enterprise workflows. The interesting part is not the model call. It is everything around it: what the workflow looked like before, what a wrong extraction does to a human later, how to keep the system usable when the model is uncertain, and how to ship this into software that already has users. AI that cannot live inside an existing process is a demo.
Technologies
Tools I actually use.
Not a proficiency chart. These are the materials the work is made from.
Primary
- Java
- Spring Boot
- REST APIs
- PostgreSQL
- SQL
Infrastructure
- AWS
- Cloudflare
- Docker
- CI/CD
- Messaging and integrations