Kacper Wikiel
Machine Learning Engineer. Warsaw, Poland.
I build machine learning systems. I'm most interested in the stuff that runs in the real world: on pipelines, on edge hardware, on your own laptop. Not just in a notebook.
The gap between what these models can do and what people actually ship is still huge. Most of my time goes into closing it.
On the side, I build my own products. I think building is the best way to think.
k.wikiel@gmail.com ·
GitHub ·
LinkedIn ·
CV (EN) ·
CV (PL)
Work
-
Built the CV pipeline that finds welds on a country's fuel pipelines. Along the way found a sensor bias nobody had noticed.
-
A Raspberry Pi that detects aircraft, tracks them with a gimbal, and computes their GPS coordinates in real time. Edge hardware is going to matter more than people think.
-
Document platform with ABAC and hash-chain audit trails for classified work. The boring infrastructure that makes serious things possible.
Things I'm Building
-
Hardparse. OCR API. Tables, formulas, charts, seals.
-
CodeSOTA. Benchmark comparison site for ML. ~30k monthly users. Progress is faster than the papers suggest.
-
TasteRay. Movie recommendations that come from a real conversation.
Experience
- PwC, ML Engineer (contract), 2025.
- National Pipeline Operator, ML / CV Engineer, 2024 to 2025.
- Procter & Gamble, Software Engineer, Generative AI, 2022 to 2024.
- Igoria Trade, Python Developer, 2021 to 2022.
Education
- Computer Science, PJATK (in progress)
- Physics, University of Warsaw
- XIV LO im. Staszica, Warsaw. IYPT 2012 laureate (International Young Physicists' Tournament).
Certifications
- Google Cloud Associate Engineer
- Azure AI Engineer