Szymon Chirowski

Open to ML internships & part-time roles

Szymon
Chirowski.

Applied Data Science & AI Student · Machine Learning · MLOps · Applied Research

I'm an Applied Data Science & AI student who likes building ML systems end to end — from a notebook experiment to something you can actually deploy, monitor and trust. MLOps is the part I keep gravitating toward.

Download CV ↓ Get in touch ↗
Szymon Chirowski, Applied Data Science & AI Student
Breda, Netherlands
3rd year Applied Data Science & AI student
01

Selected work

A mix of client work, academic projects and research — each led by what it achieved and the MLOps behind it: deployment, pipelines, monitoring and scale.

01

Academic group project · BUas Academic

Real-Time Sign Language Recognition

Built a real-time Dutch Sign Language (NGT) recognition system — from a MediaPipe + EfficientNet-B0 model to a fully deployed, authenticated inference API — wrapped in a production MLOps stack with an automated Azure ML training pipeline, a versioned model registry, CI/CD and multi-target Docker deployment.

MLOps

An end-to-end production pipeline: Azure ML training with accuracy/F1 quality gates, automatic model versioning in a registry the API pulls from on startup, GitHub Actions CI/CD, multi-target Docker deploys (Azure + on-prem via Portainer GitOps), and an enforced 90% test-coverage bar.

  • PyTorch
  • EfficientNet-B0
  • MediaPipe
  • FastAPI
  • Azure ML
  • MLflow
  • Docker
  • PostgreSQL
  • GitHub Actions
Project details
02

Breda University of Applied Sciences Client

BUas Study Choice Assistant

In development — pre-deployment QA complete, awaiting deployment sign-off

Designed and built a production RAG chatbot, commissioned by the BUas administration, that answers prospective-student questions on programmes, admissions, fees, housing and visas — reducing student-office load. A formal 26-tester QA round measured an 81.6% functional and 90.0% adversarial pass rate ahead of deployment.

Systems

A multi-service system with its own data ingestion pipeline (Selenium-based scrapers + PDF ingestion into a Chroma vector store), an LLM-supervisor moderation layer, an async PostgreSQL store with Alembic migrations, a feedback admin panel, and a Dockerised multi-container deployment.

  • Python
  • FastAPI
  • LangChain
  • ChromaDB
  • Anthropic Claude
  • PostgreSQL
  • Selenium
  • Docker
  • React
Project details
03

Netherlands Plant Eco-phenotyping Centre Client

Automated Root Phenotyping & Inoculation

Built a computer-vision and reinforcement-learning system that segments plant roots and drives a liquid-handling robot to inoculate them — reaching 0.18mm targeting accuracy and cutting time per Petri dish 21×, from ~3.5 minutes to roughly 10 seconds.

MLOps

Tracked every training run in Weights & Biases — loss curves, segmentation metrics and model comparisons — and built a baseline-first pipeline that made each modelling decision reproducible and auditable.

  • PyTorch
  • SegFormer
  • OpenCV
  • Stable-Baselines3
  • Gymnasium
  • Weights & Biases
  • Opentrons
Project details
04

Content Intelligence Agency Client

Emotion Classification in TV Content

Built an end-to-end NLP pipeline that turns raw video into timestamped emotion labels across six emotions — a local, transparent alternative to expensive cloud LLM services, with the final Transformer model reaching an F1 of 0.75.

Focus

Compared transcription engines on Word Error Rate and classical-vs-Transformer models on a common evaluation, then wired explainability (attention visualisation) into the loop to diagnose failure modes.

  • PyTorch
  • Transformers
  • BERT
  • RoBERTa
  • Whisper
  • AssemblyAI
  • scikit-learn
Project details
05

IKEA Breda Client

POS Terminal Deployment Analytics

Turned 1,300+ raw POS-terminal deployment events into a Power BI dashboard that surfaced a 92% compliance rate, flagged anomalous logging across 67 terminals, and gave store operations a clear month-over-month view of where deployments were slipping.

Focus

Modelled the full terminal lifecycle in DAX and built automated anomaly flags so operations could spot data-integrity and compliance issues without manual review.

  • Power BI
  • Power Query
  • DAX
Project details
View all 7 projects ↗
02

Experience

Where I've built and run ML in production — and the work that came out of it.

Feb 2026 — Present

Research Assistant

Breda University of Applied Sciences

Working on the Tangible Landscape project, focused on geospatial data projection and processing — including 3D scanning and point-cloud workflows that turn physical terrain models into live, analysable data.

  • 3D Scanning
  • Point Cloud Processing
  • Python
  • Geospatial Data
Apr 2025 — Present

AI Chatbot Developer · Student Assistant

Breda University of Applied Sciences

Commissioned by the university administration, I led end-to-end development of an AI chatbot that helps prospective students with program discovery, application tracking and admissions questions. I designed the inclusive, adaptive conversation flows and built the web-scraping and API-based data pipelines that keep its answers accurate against official university data — working directly with academic and administrative staff to translate their needs into features.

  • LangChain
  • Ollama
  • PostgreSQL
  • Python
Aug — Sep 2023

SOC Trainee

DAGMA Bezpieczeństwo IT · Katowice

Summer apprenticeship in a Security Operations Center. Worked with the ELK Stack (Elasticsearch, Logstash, Kibana) and SIEM tooling for security-event monitoring and threat detection, gaining hands-on experience in log analysis and incident-triage workflows.

  • ELK Stack
  • SIEM
  • Log Analysis
03

Education

2024 — 2028 (expected)

B.Sc. Applied Data Science & Artificial Intelligence

Breda University of Applied Sciences

Focus — Applied ML · Computer vision · MLOps · Production data systems

2019 — 2024

IT Technician (Technical Secondary School)

Technikum Lotnicze, Katowice

Focus — IT infrastructure · Server administration (Windows & Linux) · Programming fundamentals

04

About

I'd rather ship a model that keeps working in the real world than top a leaderboard I'll never deploy.

I'm a 3rd year Applied Data Science & AI student at Breda University of Applied Sciences. Most of what I know I picked up by building — through work, research projects and things I make on weekends.

The part I keep coming back to is MLOps: feature pipelines, CI/CD for models, monitoring and safe rollbacks. Turning a notebook experiment into something dependable is the problem I find most fun.

When I'm not studying you'll find me reading ML-systems papers, tinkering with my home-lab, and over-engineering my coffee.

  • Breda, NL
  • Data Science & AI
  • Open to internships
  • MLOps-minded