AI ENGINEER · FULL STACK DEVELOPER · COMPUTER SCIENCE STUDENT
I build intelligent systems - from model to product.
Computer science engineer specializing in AI engineering: machine learning, deep learning, NLP and vector search - grounded in real-world fullstack and mobile development.
About
I care about AI that actually ships.
I'm pursuing a Master's in Artificial Intelligence Engineering at Lublin University of Technology, after finishing my engineering degree in software engineering with a 5.0 final grade. My work spans classical machine learning, deep learning, LLMs, recommender systems and interpretable AI - always with an engineer's bias for measurable results.
Before AI, I shipped products: a production mobile app on iOS and Android, fullstack web applications and documented APIs. That foundation means my models don't stay in notebooks - they become systems people use.
TOOLBOX
AI / ML
ML models - classification, regression
Deep learning - PyTorch, TensorFlow
LLMs & prompt engineering
NLP - transformers
Vector databases & RAG
Recommender systems
Fuzzy & interpretable systems (XAI)
Languages & frameworks
Python - NumPy, pandas
TypeScript / JavaScript
Next.js, React
React Native / Expo
FastAPI, Express.js
Data & cloud
PostgreSQL, MongoDB, Prisma
Firebase, Supabase
AWS - core services, Amazon Bedrock
Qdrant - vector database
Docker
REST API design - Swagger / OpenAPI
Ways of working
Git / GitHub, CI/CD
Agile / Scrum
Cross-functional teamwork
Releases - TestFlight, Google Play Console, fastlane
Analytics - PostHog
Claude Code - AI speeds up the typing; the logic and system architecture stay mine
Research-grade AI work with published metrics, next to production-minded engineering.
Fuzzy Sleep Apnea Detector
Python · scikit-learn · fuzzy logic · XAI
Hybrid fuzzy expert system detecting sleep apnea from raw ECG. Four calibrated ML models (RF, SVM, KNN, MLP) feed a fuzzy inference engine that produces an interpretable risk score.
LightGCN Recommender
PyTorch · graph neural networks
From-scratch implementation of LightGCN (He et al., SIGIR 2020): graph-based collaborative filtering with K-layer embedding propagation and BPR loss, trained on Gowalla - 1M interactions.
Percipio
ESP32 · Next.js · Firebase · mmWave radar
End-to-end IoT room-awareness system: ESP32 firmware streams mmWave radar tracking, climate and RFID access data to Firestore in real time, visualised live in a Next.js dashboard with Telegram alerts.
PetStayTion API
Next.js · PostgreSQL · Firebase · Docker
Backend for a pet-care services platform - user, pet and service-offer management. Fully documented with Swagger / OpenAPI and containerized with Docker.
Hotel Website
HTML · CSS · JavaScript
A polished, responsive marketing site for a hotel - clean, dependency-free front-end craft.
Portfolio
Next.js · TypeScript · Tailwind CSS
This site: a bilingual (EN/PL) portfolio built on the Next.js App Router, statically exported and deployed to GitHub Pages through a GitHub Actions pipeline.
Experience
Oct 2026 - Feb 2027
IncomingAI Engineer - Schuman Trainee
European Parliament - Brussels, Belgium
Jul 2025 - present
Mobile Application Developer
diagno6 - Wrocław, Poland
Jul 2025
Fullstack Developer - apprenticeship
Transition Technologies MS - Lublin, Poland
Education & certificates
Feb 2026 - present
MSc Eng, Computer Science
Specialization: Artificial Intelligence Engineering - Lublin University of Technology
Oct 2022 - Feb 2026
BEng, Computer Science - 5.0
Specialization: Software Engineering - Lublin University of Technology
CERTIFICATES
Transformer-Based Natural Language Processing
NVIDIA
2026
AI in Business Development
2025
LanguageCert Test of English - C1
LanguageCert
2023
LANGUAGES
Polish - native · English - C1 · German - B1 · French - learning
GET IN TOUCH
Let's build something intelligent.
Open to AI engineering opportunities, collaboration on interesting problems, and good conversations about ML.
jacek.kozlovski@gmail.com