Production-grade AI products, from idea to operation.
Twelve years of backend and distributed systems, now applied to LLM agents. I cover the whole cycle: product decisions, architecture, code, release and operation.
Ownership across the full product cycle
A product with AI at its core depends on understanding the model, the system around it and the business behind it. I work across all three, from idea to production.
- 01
Architect
Service boundaries, data models and integration contracts designed for change, with trade-offs written down.
- 02
Release & operate
Automated delivery, Kubernetes deployments, monitoring and incident response. I stay responsible for what goes live.
- 03
Measure & improve
Quality and reliability tracked with real metrics and evaluation, so every change is verified, not assumed.
Solving problems, whatever the language
A problem solver, not a code writer
I start from the outcome: what has to be true for the product to work, and what could break it. Code is one of the means, never the goal.
Languages are tools
Hands-on with Java, Python and Rust, plus TypeScript. With AI agents writing much of the code, the language matters less than knowing how to structure, review and verify the work.
Judgment over typing
I direct AI coding agents, set the guardrails and own the review. Speed comes from delegation; correctness comes from knowing what good looks like.
What I can build and run for you
AI agents that take action
Tool-using agents on LangGraph and MCP with planning, state, retries and fallbacks, evaluation and cost control, built to be audited when something goes wrong.
Production RAG
Retrieval pipelines from ingestion to ranking, with answer quality measured on real data.
Scalable backend
Java and Spring Boot services with Kafka and PostgreSQL: secure APIs, clear data ownership and event-driven integration.
Platform and operations
Containers and Kubernetes in daily use: delivery pipelines, configuration, secrets, observability and incident response.
Twelve years shipping production systems
- Dec 2022 — PresentSinchSenior Java Developer
- Built an MCP server that exposes multiple internal OpenAPIs to AI agents, including secure email with attachments.
- Designed a multi-agent orchestration system with tools for knowledge-base management and browser-based web automation.
- Led the core platform upgrade to Spring Boot 3 and Java 21 and improved high-load, time-sensitive processing.
- May 2021 — Nov 2022AptivSenior Software Engineer
- Serverless microservices for device provisioning on AWS Lambda and Azure Functions.
- Secure ETL pipeline with Azure Key Vault encryption.
- Sep 2018 — Apr 2021UBSPrincipal Software Engineer
- Led an MVP from a blockchain / DLT proof of concept to an enterprise platform release.
- Feb 2016 — Aug 2018Luxoft (Cisco project)Java Technical Lead
- Managed 10 developers across backend and frontend.
- Designed failover handling and a high-load data processing system on Akka.
- Nov 2013 — Feb 2016Sberbank-Technology, IQubeJava Technical Lead
- Built and led the senior team behind a core frontend microservice; designed a message verification layer with two-factor authentication.
- Delivered backend microservices from scratch to production.
Earlier: Java developer at EPAM Systems, 2012 — 2013.