n8n-automation

Hub d'automatisations personnelles

← Annuaire

Jelle Scholtalbers

AI & Full Stack Engineer - Python/Django, DevOps

Voir sur Malt

Saint-Marcellin, FR · 720€/j · Disponibilité confirmée

Artificial Intelligence (AI) Python Django REST APIs Data Management Workflow automation SAAS Gitlab CI/CD HPC Slurm n8n AI Agent FastAPI SQL PostgreSQL pgvector LLM AI Orchestration Life Sciences Snakemake Nextflow Galaxy RabbitMQ Redis Memcached Docker Kubernetes DevOps Linux Typescript React.js Grafana Prometheus Vue.js Django Rest Framework LangGraph Helm Ansible RAG Model Context Protocol LIMS SQLAlchemy Bash OpenTelemetry laboratoire

Missions

LabRise Consulting Django

janvier 2024 - Aujourd'hui(2 ans et 9 mois)

CONSEIL & AUDIT

Own products and platform work, alongside client engagements.LabID SaaS. My own multi-tenant variant of the open source lab data management and ELN platform, in active development and close to launch. Tenant isolation, onboarding and the operating model, including the cost decisions around managed services, are mine alone, unlike the open source platform where I am one of several core developers. Python, Django, DRF, PostgreSQL, Vue 3, Docker, GitLab CI/CD.Po11y. Open source status dashboard and observability stack for n8n, MIT licensed and public on GitHub. One process polls n8n's REST API read only into SQLite and serves a live execution feed, a workflow map whose edges follow real references, prewired Grafana dashboards, a watchdog, optional Prometheus and Alertmanager, opt in OpenTelemetry tracing and a read only MCP endpoint. Compose and Kubernetes. It attaches to a running production n8n without ever writing to it.FastOrchestra. Full stack pipeline over ranked job ads. React 19 and TypeScript on the front, FastAPI, SQLAlchemy 2 and PostgreSQL with pgvector behind it, API and worker in one image, Alembic migrations, Prometheus metrics with alert rules, a Helm chart installed against a kind cluster in CI, releases tagged to GHCR. Embedding runs locally on CPU, only the drafting prompt leaves the host, every model call is traced with tokens and cost. The same API speaks MCP, so an assistant drives the whole loop.AI assisted development. Self hosted LangGraph and Claude Code agents carry a tracked GitLab issue through plan, code, a self fixing test loop, merge request and automated review. Docker on self hosted runners, outbound only, with provider routing through a self hosted OmniRoute gateway, so the model is a configuration choice. The data boundary is stated plainly: prompt context including source code reaches the model provider, everything else stays on own infrastructure.Available for backend, platform and AI engineering, remote from France.

LLM Orchestration Python/Django FastAPI MCP DevOps

EMBL (European Molecular Biology Laboratory)

janvier 2024 - décembre 2025(1 an et 11 mois)

CENTRES DE RECHERCHE

LabID‑PROV (OSCARS, Horizon Europe grant No. 101129751): extending LabID to track and share data provenance with RO‑Crate, in line with FAIR principlesModeling workflows and workflow runs in LabID and linking to Life Science research infrastructure tools, integrating WorkflowHub and GalaxyEarlier contract: implemented S3 cloud storage support for LabID datasets: user‑supplied S3 credentials, URL‑referenced data and the full file management lifecycle (upload, download, deletion, sharing, metadata), compatible with EMBL and AWS S3Designed the new "Workflow Run" assay type capturing input datasets, workflow version, parameters and execution configuration from Nextflow, Snakemake and Galaxy; extended the core assay model to accept datasets (not only samples) as inputs

Cloud LIMS Fullstack Gestion de données de recherche Modélisation des données

EMBL Django

octobre 2014 - septembre 2023(8 ans et 11 mois)

CENTRES DE RECHERCHE

Lead developer of LabID lab data management and Electronic Lab Notebook platform, used by 200+ employees; led feature and technical‑ requirement decisions and ran development and production deployments via CI/CD on EMBL's managed Kubernetes clusterProduction‑hardened and maintained EMBL GitLab: configuration in version control, automated backups, Docker registry and LFS; around 2280 users, 65 groups and 3600+ projects with no major downtime or security incidentsInitiated and launched EMBL Chat (Mattermost) in 2017 as a data‑protected alternative to externally hosted Slack, growing past 1000 registered usersProvided and maintained compute services (Galaxy, RStudio) and supported and trained EMBL staff

Kubernetes Tech Lead LIMS Gitlab CI/CD Python/Django