Filip M.

Senior Java backend engineer (11+ yrs) building Kafka-based, event-driven systems that run in production: Spring Boot, Apache Kafka, Apache Flink and AWS, for healthcare, fintech and logistics. I currently own production readiness for a US healthcare company's real-time clinical data pipeline (~30 microservices, ~50 Kafka topics on Amazon MSK). I led a Flink de-duplication rollout with exactly-once delivery that removes 8–12% of vitals traffic (~1.7M redundant records/day), and built the monitoring, runbooks and recovery tooling around it. What I can take off your plate: • Spring Boot (Java 17/21) APIs and microservices: design, build, ship • Apache Kafka: event-driven architecture, Avro/Schema Registry, exactly-once, lag and reliability issues • Stream processing with Apache Flink (incl. Amazon Managed Service for Apache Flink) • AWS backends (ECS, Lambda, MSK, DynamoDB, S3) provisioned with Terraform • Production hardening: Datadog dashboards and alerts, runbooks, failure testing, recovery paths • PostgreSQL data modeling and performance; auth and multi-tenant security (Spring Security, JWT, RBAC/ABAC) • Payments: Stripe Certified Professional Developer • Python/Django when the project needs it Selected work: • Healthcare (US): real-time HL7 v2 clinical pipeline, Flink de-dup, Lambda-based Kafka replay/recovery tool, 13 prioritized Datadog monitors • FIS: post-trade derivatives clearing with CQRS and event sourcing (Akka/Lagom, Kafka, Couchbase) • HTEC: 20+ microservice platform for autonomous-vehicle operations (dispatch, routing, tracking, billing) • End-to-end builds: multi-tenant SaaS backend; website monitoring platform (Django, Celery, Stripe) Good fit if you need: • A new Spring Boot backend or MVP taken to production • Kafka introduced, fixed or scaled: duplicates, lost messages, lag, schema evolution • A monolith split into services, or an existing system made observable and reliable • A senior engineer for architecture or code review On Upwork: 11 completed jobs rated 4.8–5.0, 750+ hours. EU-based with solid US overlap. Send me the problem, stack and timeline, and I'll reply with a concrete approach and estimate.