Founder at t3chnico.com (2024-11 – Present)
- Building production AI products end to end for clients across Belgium and the EU, including agentic pipelines, RAG systems, and full-stack SaaS. I own scope, delivery, and the running system, with a direct client line.
- Continuous R&D on open-source tooling around multi-agent orchestration, developer infrastructure, LLM knowledge systems and Software Factories
- Preferred stack: TypeScript, Turborepo, Hono with oRPC, TanStack Router, Drizzle with Postgres, pgvector, Better Auth, Vercel AI SDK/ Mastra for AI. Hetzner, Proxmox, Docker. Tauri, React Native. Python FastAPI for microservices.
Founding Engineer at askaristotl.ai (2025-08 – 2026-02)
- First engineering hire. Helped build the new AI native product MVP from scratch
- Made a custom document parser FastAPI microservice, that outperformed existing FOSS solutions and most paid solutions by leveraging SOTA VLMs and an elegant prompt engineering trick. 10x+ cost savings over paid solutions with comparable performance. Way faster then most paid/FOSS solutions as well. As all pages of e.g. a 1000 page document were processed in parallel, the only real bottleneck was provider rate limits (we were using open models through OpenRouter). Most 200 page documents were processed under a minute.
- Learned valuable lessons about startups and entrepreneurship: validate before you build, and anchor every product decision to a clear go-to-market strategy from day one.
- During this time, Spec Driven Development really became the norm. We experimented and learned a lot about what agentic coding workflows work in a real team environment.
- Stack: fullstack Next.js, Vercel, Supabase, Railway. Then we migrated away to Next.js frontend, dedicated Hono backend, Redis, Turborepo, AWS, Terraform
AI Focused Python Backend Developer at usewhale.io (2024-07 – 2025-03)
- Was responsible for all AI features in the platform
- Ideated one of their top performing features: Video To SOP. It imports videos of physical/digital processes into the platform, transforms it into textual step-by-step instructions with screenshots at key moments, RAGable by the chatbot
- Did the smart chunking strategy of the document import feature
- Set up Langfuse, LiteLLM, evaluation pipelines
- Improved the chatbot, added tools, made it more reliable
- Implemented Google Drive doc import into the platform in 3 weeks. Tested many frameworks. Concluded custom build with Drive API v3 was the right approach. Files got synced, chunked, embedded and made queryable by the chatbot
- After 9 months, I was offered a team lead role to start a new AI team within the company
- Stack: Django DRF, GCP, Postgres, Kubernetes
Full Stack Developer at ProFunctional (2023-02 – 2023-05)
- Created a reservation system for 30+ padel clubs with secure keypad access
- Integrates with an API for code generation and controls court lighting based on sunset/sunrise
- Club owners have a dashboard to view metrics and adjust settings
Junior Software Developer - Automation at ST Engineering (Former Newtec) (2020-10 – 2020-10)
- Developed a python script that automated log extraction from modems that were multiple server hops away
- Manually this task took up to 15 minutes, the script dropped this to 1 minute or less.
- I learned what the key requirements are to save time & effort through scripting & automation
Summer Job Machine Learning - R&D project at ST Engineering (Former Newtec) (2020-07 – 2020-08)
- Using Machine Learning to detect anomalies in log files
- I learned that rather than reinventing the wheel, implementing & elaborating an existing solution was the right choice to save time, effort and money.
- Python, Flask, bootstrap, Github, ElasticSearch, Grafana
