Omschrijving

If you need AI integrated into your product or a full-stack application built from scratch, I can take it from idea to production.

I am a software developer and AI engineer based in Amsterdam with an MSc in Artificial Intelligence from VU Amsterdam. Over the past 3 years I have worked professionally on production AI systems: LLM-powered applications, RAG pipelines, and agentic infrastructure that real users depend on.

I have shipped systems in a health data context where accuracy is critical, and built data pipelines in a regulated financial environment where correctness is not optional. Both taught me that the hardest part of AI engineering is not the model but everything around it: the infrastructure, the observability, the edge cases.

Beyond client work I build open-source developer tooling in my own time. That reflects how I approach problems generally: I go deep, solve things properly, and build for production rather than demos.

My stack covers Python, TypeScript, React, Rust and Go, with strong experience in LLM APIs, agent frameworks, REST APIs, and full-stack web development. I work across the entire product, from designing interfaces to wiring up backend services and deploying infrastructure.

I take on AI integration projects, full-stack web applications, agentic systems, and API development. I work remotely and communicate in English, Romanian, Dutch and German.

Talen

Roemeens
Tweetalig / moedertaal
Engels
Tweetalig / moedertaal
Duits
Beperkte professionele capaciteit
Nederlands
Beperkte professionele capaciteit

Voorkeuren voor werkplek

Uitsluitend remote

Werkt voornamelijk remote

LongevAI
AI & Software Developer
MEDISCHE SECTOR
februari 2025 - Vandaag (1 jaar en 4 maanden)
Amsterdam, Nederland
• • LongevOS: architected a multi-tenant SaaS clinical operating system (FastAPI, Next.js, MongoDB) that turns consultations and lab PDFs into clinician-approved health plans through a node-based LangGraph orchestration layer integrated with Vertex AI, deployed as a multi-region containerized service on Koyeb.
• • DNA-analysis platform: built a B2C genomics product (Next.js, FastAPI, PostgreSQL) whose extraction pipeline auto-detects and parses multi-format raw DNA and enriches markers against GWAS Catalog, ClinVar and PharmGKB via a batched variant-annotation API, processing 1,000 markers in 10 to 30 seconds.
• • B2B lead-gen engine: engineered an 8-stage LangGraph discovery pipeline with a self-correcting refinement loop and resilient multi-provider search (Serper/Tavily/Brave) behind circuit-breaker failover, persisted to MongoDB Atlas.
• • Consultation platform: built an end-to-end application for multi-user clinics covering user and client management, live recording with real-time transcription, Celery-worker AI processing, and an AI copilot for editing and auto-generating versioned reports and prescriptions, on FastAPI with WebSockets, MongoDB, and Vertex AI.
• • Research newsletter pipeline: built an automated content engine ingesting longevity research via RSS/web scraping (feedparser, BeautifulSoup), triaging and summarizing it with multiple LLMs (OpenAI/Anthropic/Gemini), and drafting curated newsletters (FastAPI, Celery).
Python Typescript AI Automation Next.js LLM
Van Lanschot Kempen
Fullstack Developer
BANKEN & VERZEKERINGEN
september 2024 - Vandaag (1 jaar en 9 maanden)
Amsterdam, Nederland
• • Communication-archiving platform: engineered a regulated multi-region archive (.NET 10/ASP.NET Core, Angular 20, Cosmos DB, Azure AI Search) ingesting voice, email, and chat from multiple source systems via 15+ Azure Functions (Python and .NET), with WORM-backed retention, dual-region failover, and a Cosmos fallback search tier for archived data.
• • Data-governance platform: built a full-stack metadata and access-governance system (Flask, SQLAlchemy, React/TypeScript/Redux Toolkit, Material UI, D3.js lineage graphs) with Celery/RabbitMQ/Redis async pipelines, a Git-backed data-model editor with branch-based approval workflows via GitHub App webhooks, and a Databricks-powered RAG + NL2SQL copilot streamed over SSE.
• • Provisioned both platforms as containerized microservices on Azure Container Apps and App Service via Bicep IaC, with Azure DevOps CI/CD across a four-stage DTAP pipeline, Managed Identity for secret-less Azure auth, private-endpoint networking, and Application Insights/Log Analytics observability.
Microsoft Azure Azure DevOps .NET C# React.js
Neolocus
Software Developer
ARCHITECTUUR & STEDENBOUWKUNDE
juli 2024 - september 2024 (2 maanden)
Amsterdam, Nederland
• • Engineered the core AI rendering pipeline with ControlNet and Stable Diffusion, turning sketches/photos and prompts into photorealistic interiors, plus a Segment Anything (SAM) + inpainting pipeline for furniture removal/replacement.
• • Delivered generative-AI features that helped drive the company to $1.7M in revenue and a strategic equity investment from Saudi real-estate giant ROSHN.
Python Machine learning Computer Vision GenerativeAI AI Automation