2024
2023
2022
2022.12
Started, among others – Polish ML Community https://polishml.community
Gave a talk
2022.11
2022.01
Won a hoodie
2021
2021.12
Linkedin 24 posts series:
- https://www.linkedin.com/pulse/do-czego-sieci-neuronowe-w-analizie-obrazu-karol-majek-phd/
- https://www.linkedin.com/pulse/od-zbioru-danych-do-pierwszej-sieci-neuronowej-karol-majek-phd/
- https://www.linkedin.com/pulse/yaxb-w-kerasie-karol-majek-phd/
- https://www.linkedin.com/pulse/trening-w-uczeniu-nadzorowanym-karol-majek-phd/
- https://www.linkedin.com/pulse/overfitting-nadmierne-dopasowanie-karol-majek-phd/
- https://www.linkedin.com/pulse/jak-oszuka%25C4%2587-overfitting-karol-majek-phd
- https://www.linkedin.com/pulse/transfer-learning-zaoszcz%25C4%2599dzi-tw%25C3%25B3j-czas-karol-majek-phd
- https://www.linkedin.com/pulse/data-augmentation-karol-majek-phd/
- https://www.linkedin.com/pulse/hiperparametry-vs-parametry-karol-majek-phd/
- https://www.linkedin.com/pulse/gradient-hessian-i-predykcja-wag-karol-majek-phd/
- https://www.linkedin.com/pulse/historia-architektur-cnn-karol-majek-phd/
- https://www.linkedin.com/pulse/frameworki-karol-majek-phd/
- https://www.linkedin.com/pulse/koniec-prac-nad-wykrywaniem-obiekt%25C3%25B3w-karol-majek-phd
- https://www.linkedin.com/pulse/dlaczego-tak-karol-majek-phd/
- https://www.linkedin.com/pulse/poprawa-klasyfikacji-obrazu-karol-majek-phd
- https://www.linkedin.com/pulse/du%25C5%25BCo-zdj%25C4%2599%25C4%2587-ma%25C5%2582o-etykiet-co-robi%25C4%2587-karol-majek-phd/
- https://www.linkedin.com/pulse/pytorch-image-models-karol-majek-phd/
- https://www.linkedin.com/pulse/modele-jako-aplikacje-demo-karol-majek-phd/
- https://www.linkedin.com/pulse/czy-klasyfikacja-obrazu-jest-rozwi%25C4%2585zana-karol-majek-phd/
- https://www.linkedin.com/pulse/preprocessing-kontra-przestrzenie-barw-karol-majek-phd/
- https://www.linkedin.com/pulse/cnn-vs-mlp-karol-majek-phd/
- https://www.linkedin.com/pulse/konwolucja-vs-transformer-karol-majek-phd
- https://www.linkedin.com/pulse/obrazy-vs-3d-karol-majek-phd/
- https://www.linkedin.com/pulse/udost%25C4%2599pnianie-wdra%25C5%25BCanie-modeli-karol-majek-phd
2021.06
Program Szkoleniowy Klasyfikacja Obrazu – https://deepdrive.pl/klasyfikacja
2020
Waste detection – Mentor
2019
2019.11
PhD defence
2019.07
ERL Emergency 2019
https://sites.google.com/site/erlemergency2019laspezia/home
Enrich 2019
2019.05
WSDC 5
https://deepdrive.pl/wsdc-5-car-accident-risk-prediction-rc-top-down-view-from-side-cameras/
Workshop – TensorFlow Object Detection API
https://nlpday.pl
2019.04
Data Science Olsztyn #3: Object detection
https://evenea.pl/imprezy/szkolenia/olsztyn/data-science-olsztyn-3-218660/
2019.03
2019.02
WSDC 4
https://deepdrive.pl/wsdc-4-traffic-management-and-sdc-datasets/
2018
2018.12
PL in ML – Deep Learning for Self-Driving Cars (6h workshop)
https://conference.mlinpl.org/archive/2018/
Workshop – Deep Learning for Self-Driving Cars
https://conference.mlinpl.org/archive/2018/
2018.11
Autonomous Racing – from the simulator to a real track.
In this talk, Karol will present lessons learned from two autonomous driving events Self-Racing Cars 2017 and 2018. He participated in the first event in a team supported by the Udacity with an End-to-End network trained in a simulator. In the 2nd event, the team used semantic segmentation and path planning to improve the results.
Autonomous Racing using Deep Neural Networks – HIPERT Lab, UNIMORE
2018.10
F1/10 Challenge – autonomous 1:10 racing; Turin, Italy
1st place!
GDG Radzymin #2
- Image analysis in public clouds: AWS, GCP, Azure
Data Workshop Club Conf:
- Workshop: How to use RL in the browser to train a car to drive autonomously – using metacar.js
- Talk: My road to object detection
2018.09
ELROB 2018 robotic competition
http://www.elrob.org/elrob-2018
- 1st place in Reconnoitring of Structures
- Best Scientific Solution award
Testing Magic Leap
2018.08
ROS2 experiments
- on Linux, Windows, communication, ROS1 bridge
Testing LeGO-LOAM
2018.07 – DeepFakes
2018.06
Igrzyska chmury (Cloud Games!)
Public Cloud Competition – 5th place
Multi-Cloud
Igrzyska chmury [PL]
No odometry mapping using Hector SLAM
F1Tenth
Dataset recorded – more info [PL]
2018.05
Hackart
1st place in Hackathon MNW [PL]
Team:
2018.03
Self-Racing Cars 2018 – Team Soulless
2017
2017.10-12 – FKIE – Photorealistic 3D mapping
2017 Enrich 2017
Interactive DEMO in the browser!
2017.04 – Self-Racing Cars – Udacity Team Soulless
2016
2016.11 – Image-Based Localization – Udacity Challenge 3
Localize the car using only a single camera.
2015
- Würzburg, Germany, German Academic Exchange Service (DAAD), Sensor data processing using General Purpose Computations on GPUs
- Polytech Nantes, France, Development of a multimedia library for audio and video services based on hardware accelerator Tegra K1
- Testing a mobile system for supporting spatial design
- Integrated Components for Assisted Rescue and Unmanned Search operations (ICARUS)
- Toolbox Implementation for Removal of Anti-personal Mines and Submunitions (TIRAMISU)
- EURATHLON 2015 – European Outdoor Robotics Competition, ICARUS team
2013
- EURATHLON 2013 Land Competition, IMM-IAIR team
- DARPA Virtual Robotics Challenge, Robot-nicy team, VRC finalist