Hello, I’m Filip — building robust, elegant software.

Cześć, jestem Filip — tworzę solidne i eleganckie oprogramowanie.

Salut, je suis Filip — je crée des logiciels robustes et élégants.

I focus on clean architecture and measurable impact of implemented solutions.

Last GitHub activity: on
Portrait of Filip Jakubowski

About

Working on performant systems and clear UX. Passionate about architecture, data, and shipping fast without breaking quality.

Projects

Project 1 preview

ASPDNet, RSOC Building Patch and Inference Web App

Patched and productionized version of the original ASPDNet framework designed for building counting in aerial imagery. The code is adapted to PyTorch 2.5.1, Python 3.9, and CUDA 12.3, and validated on the RSOC dataset.

  • PyTorch
  • CUDA
  • Python3
Project 2 preview

Military Vehicle Classification on MSTAR Imagery

How easy is it to train and apply object classification for the MSTAR IU Mixed Targets dataset using YOLOv8? I think, I've found out.

  • Python
  • Ultralytics
  • Classification
Project 3 preview

Corn Plant Segmentation and Tracking for Agri-Robot

Training an Instance Segmentation model with tracking for an agricultural robot.

  • OpenCV
  • Instance Segmentation
  • Ultralytics
Project 4 preview

Cybersecurity Attack Prediction

A data engineering and machine learning pipeline designed to predict potential cybersecurity attacks from raw log data. The project focuses on building a robust preprocessing and feature engineering workflow, enabling multiple models (neural networks, tree-based, ensemble) to be trained and evaluated efficiently.

  • Feature Engineering
  • PCA
  • Model Training and Evaluation

Agent-Based Modeling — Work in Progress

A work-in-progress project using NetLogo to simulate complex systems through agent-based modeling. The focus is on emergent behavior analysis and scenario testing.

  • NetLogo
  • Simulation
  • ABM

Ontology for Mapping and Detecting Information Operations — Work in Progress

A work-in-progress proposal for an ontology designed to structure and map data for the detection of coordinated information operations online.

  • Ontology
  • SPARQL
  • Knowledge Graph

Skills

Languages & Frameworks

  • Python
  • C++
  • SPARQL
  • PyTorch
  • Bash
  • NumPy
  • Pandas
  • HTML / CSS
  • RDFS

Tools & DevOps

  • AWS
  • Docker
  • Git
  • Google Cloud
  • GraphDB
  • Linux
  • OpenCV
  • TensorBoard

Contact

Email 1: jakubowski.filip@outlook.com

Email 2: filip.jakubowski@edu.dsti.institute

Phone 1: +48 795 428 863

Phone 2: +33 666 563 790