|
I build data systems that scale — from raw ingestion to business-ready models. My focus is the modern data stack: reliable ELT on Snowflake, clean and tested dbt models, and semantic layers that give teams one source of truth. I'm now extending that foundation into trustworthy AI — my MSc thesis tackles uncertainty calibration in retrieval-augmented LLMs: whether a model's confidence can be trusted for risk-sensitive decisions, and how embedding quality shapes that reliability (LLM-as-judge, retrieval metrics, LangChain & LangGraph).
|
|
Robust ELT pipelines on Snowflake — tested, monitored, and built to be trusted in production. |
Clean, documented dbt models and semantic layers that give analysts a single source of truth. |
RAG pipelines with LangChain, Chroma & FAISS — rigorously evaluated with LLM-as-judge and retrieval metrics (MRR, nDCG, recall, precision). |
| 🏗️ Data Engineering & Analytics |
|
| 📐 Data Modeling |
|
| 🤖 AI / GenAI |
|
| 📊 BI & Tooling |
|
| Project | Stack | Status | |
|---|---|---|---|
| 01 | RAG Uncertainty Calibration — MSc thesis | RAG · FAISS · scikit-learn |
✅ Live |
| 02 | Enterprise Retail Data Mesh — dbt · Terraform · Cortex AI | Snowflake · dbt · Terraform |
✅ Live |
| 03 | Operations Copilot — NL→SQL analytics | Snowflake Cortex · Streamlit |
✅ Live |
BGE beat OpenAI embeddings · isotonic calibration cut ECE ~32% — see the thesis repo.
I'm open to Data Engineering · Analytics Engineering · AI Engineer roles across Europe. The fastest way to reach me:
Thanks for stopping by — let's build something reliable.