Camunda enterprise workflow automation platform
I led the architecture and development of an enterprise workflow orchestration platform built around Camunda and BPMN, designed to automate complex, mission-critical business processes.
I developed scalable backend services using Node.js/NestJS and Python/FastAPI, with React and Next.js powering process management and analytics experiences. I also integrated OpenAI and LangChain to introduce AI-driven decision capabilities, enabling workflows to interpret documents, classify incoming data, make contextual decisions, and automatically trigger actions across connected services.
To support high-volume workflow execution, I designed a distributed, event-driven processing architecture using queue-based systems, reducing workflow latency by 35%. I optimized GraphQL and REST APIs with Redis caching and asynchronous processing, improving API response times by 30%.
I also built interactive process analytics dashboards that gave teams visibility into workflow status, execution metrics, failures, and bottlenecks, contributing to a 40% improvement in operational efficiency.
The platform was containerized with Docker and deployed through Kubernetes-based CI/CD pipelines, reducing release cycle time by 25%.
I built the back office that runs my whole business.
One page: money in, money owed, money overdue, money spent. Every brand on its own line. Six months of cash flow. A button that hides all the figures, so I can open it in a meeting without showing my revenue.
Before this, everything lived in spreadsheets, emails and my head, and checking the month took an afternoon of adding up by hand.
It was the first big thing I built with AI. It's still what I use every day, and it's what I now build for other businesses.