A selection of projects I've worked on; from solar performance tracking and attribution models to data pipelines and LLM cost optimization at scale. For how I approach this work, see the about page.
Four years of full-stack data ownership at a Dutch solar company; from inverter and battery tracking to lead attribution.
Built the operational tracking systems for a national solar portfolio: inverter uptime and generation, plus monitoring for Enphase battery systems, later extended to SolarEdge inverters and batteries. Also built a fully in-house lead attribution model from the ground up; data engineering (PostgreSQL, Python, MS Fabric), analytics layer, and business logic. Both carried marketing and business decisions through 4 years of market volatility.
Optimized LLM pipelines for a healthcare AI startup; same quality, lower cost, switchable models.
Optimized prompt pipelines at scale. Built a framework for switching between AI models using golden datasets; without sacrificing output quality. Significantly reduced per-request costs for Reg4U's Declaratiescan.ai product.
Built dashboards and integrations for better marketing attribution.
Co-built a real-time call transcription platform with AI summarization, sentiment analysis, and a transcript chatbot.
Co-founded and co-built. Transcribed customer and sales calls in near real-time. Added AI summarization of key points and next steps, sentiment analysis for red flags and signals, and a ChatGPT-style chatbot for querying transcripts.
Designed a data modeling framework for domain datasets at one of the Netherlands' largest banks.
Designed a structured data modeling framework for domain datasets. Supported data governance across teams and worked on data mapping for large-scale internal datasets.
Data engineering on GCP; pipelines, PostgreSQL, cloud infrastructure.
Data engineering work on Google Cloud Platform. Pipeline development, PostgreSQL, and cloud infrastructure.
Want to talk? Get in touch →