Senior Full Stack Developer | Consultant | DevOps | Life Science | AI engineer

Backend-focused full-stack and AI engineer with a scientific computing background, specializing in data-intensive, research-driven, and AI-enabled applications.

My main work has been on the full stack development and deployment of LabID, an open source data management platform for laboratories: technical lead at EMBL, today one of its core developers. In parallel I build its multi-tenant SaaS variant (django-tenants, Authentik, S3). Working directly with users throughout my career allows me to turn feature requests into clear specifications and see the implications on the system beyond the necessary code changes.

I can start new projects from scratch and work out the specifications with you or your team, or dive into existing and legacy software. Engagements run end-to-end, from user requirements through architecture, implementation, and production deployment.

Further technical capabilities:

AI engineering. LLM capabilities in existing products and workflows: chat assistants grounded in business context, retrieval with local embeddings and pgvector, structured extraction from free text, evaluation against labelled data, tracing in Langfuse. I also build the AI-assisted development tooling itself: self-hosted agents (LangGraph + Claude Code headless, MCP servers) that take a tracked issue through plan, self-fixing test loop, merge request and automated review. They run as Docker containers on my own host, outbound-only, with OmniRoute routing to Claude, OpenAI, or a local model when code must stay in-house. A fit for regulated, data-resident, public-sector work.

Backend. Python: Django + PostgreSQL with REST APIs, and FastAPI with async SQLAlchemy 2. Redis, RabbitMQ and Celery for caching and asynchronous jobs. Authentication (JWT, OAuth, OIDC) and production concerns such as backups and scheduled jobs are part of the standard delivery.

Frontend. Vue 3 SPAs, React with TypeScript, or server-rendered Django templates with Tailwind and HTMX, tuned for complex workflows.

Deployment & infrastructure. Docker on plain VMs or Kubernetes, Helm charts, and GitLab CI/CD golden-path pipelines: versioned components, SAST, secret detection and Trivy gates, and image digests promoted from staging to production.

Code quality runs through everything: type hints, pytest against a real PostgreSQL, Playwright, pre-commit hooks, and clear documentation (MkDocs).

Adjacent experience. HPC-based scientific workflows, GitLab administration for a large research institute.

Training & enablement. Hands-on sessions on Bash, Git, CI/CD, Snakemake, and Galaxy, leaving teams self-sufficient after project delivery.

Most useful when you need a senior generalist who can translate customer requirements into working solutions and own them end-to-end, integrate modern AI capabilities into existing products, and speak the language of the scientists you're building for.