Jakub Nowicki
Software and Machine Learning Engineer
Email·CV·GitHub·LinkedIn·Google Scholar
I am a Computer Science and Applied Mathematics student at the University of Evansville, where I also compete as a Division 1 athlete. I like building things that can see, learn, and make decisions on their own, from a system that spots car accidents in real time to traffic lights that get smarter the more they run. I published my accident detection research in IEEE, and I am currently looking for an internship or job opportunity for 2026 or 2027.
Research Interests
My research interests lie in computer vision and reinforcement learning, especially their use in real world safety and infrastructure problems. I am also interested in camera based sensing of vital signs, and in how well those methods hold up across different skin tones. I am drawn to systems that can perceive their surroundings and improve their decisions over time, whether that means catching a hazard on the road as it happens or helping traffic move more efficiently through a city.
Currently Working On
For my senior project, I am building a system that estimates heart rate, heart rate variability, and breathing rate from an ordinary webcam. It combines two signals: the tiny color changes in the skin caused by each heartbeat (rPPG) and the tiny head movements caused by the pulse (rBCG). Camera based pulse estimation is known to work less well on darker skin tones, so a big part of the project is measuring how large that gap is. This is a research project, not a medical device, and I am aiming to publish the results at IEEE.
Python, OpenCV, MediaPipe, Signal Processing
Experience
I spent a summer building web interfaces for Ruby Care's AI powered telehealth platform, which serves cancer patients and their families. I shipped patient onboarding and appointment scheduling features that connect users with nutrition, mindfulness, and mental health specialists through secure video consultations, along with health tracking and journaling tools so patients could log their treatment progress and review their medical history.
Projects
A system that watches traffic and spots car accidents as they happen, using YOLO for object detection. I published the research behind it in IEEE.
YOLO, Python, OpenCV, TensorFlow
A reinforcement learning agent that learns to play Balatro, a roguelike poker card game. I trained it in two stages, first copying a hand coded strategy through behavior cloning, then improving through real game rewards, and it now wins about seventy five percent of its matches against the live game.
Python, Reinforcement Learning
A reinforcement learning system that helps traffic signals manage intersections more efficiently, built with sustainable cities in mind.
SUMO RL, Python, Q Learning
Skills
Programming: Python, JavaScript, TypeScript, C++, SQL, HTML/CSS
Machine Learning: TensorFlow, PyTorch, YOLO, OpenCV, MediaPipe, Gymnasium
Web Development: React, Next.js, Flask, REST APIs
Data & Infrastructure: PostgreSQL, NumPy, Pandas, Linux, Git
Research Areas: Reinforcement Learning, Computer Vision