Edge AI | Real-Time Detection, OCR & Classification

Engineered a real-time Automatic License Plate Recognition (ALPR) engine on the NVIDIA Jetson Orin Nano, achieving 75+ FPS at 720p while processing frames, running AI inference and encoding dual video streams. The system uses a custom GStreamer/FFmpeg pipeline to ingest and process video, running INT8 TensorRT-accelerated models for detection and OCR, all while maintaining sub-second end-to-end latency.

Real-Time Object Detection | OCR

Built an end-to-end ALPR pipeline in Python that detects, reads and redacts license plates on commodity hardware. The system operates on an asynchronous Producer-Consumer model, orchestrating the flow of data between I/O operations and AI inference Trained object detection models on a 10,000-image dataset, with best performing models achieving a mAP50 of 97.3% and mAP50-95 of 71.2%.

Machine Learning | Multi-Class Classification

Developed and fine-tuned CNN-based classifiers using pre-trained models like EfficientNetB3, ResNet50, and MobileNetV2 to classify 80 types of military aircraft from images. Created a custom data generator and applied image preprocessing to ensure compatibility. Achieved up to 95% training accuracy and 97% test accuracy with EfficientNetB3 as a base model. Visualizations of how the different models perform are available on GitHub.

Machine Learning | Multi-Class Classification

Developed CNN and RNN models to classify ECG heartbeats as normal or myocardial infarction using time-series data. Implemented a data generator for batching and applied one-hot encoding to ensure compatibility with softmax. Achieved 88+% test accuracy through hyperparameter tuning and architecture experiments. Experiment visualizations are avaialble on GitHub.

Machine Learning | Binary Classification

Developed a feed-forwad neural network to predict thyroid cancer recurrence using numerical and categorical features. Applied encoding and normalization for optimal input processing. Achieved 95+% test accuracy through hyperparameter tuning. Experiment visualizations are avaialble on GitHub.

Web Development

My animated and responsive portfolio website, built with NextJS, React and Tailwind CSS written in TypeScript. The site allows users to view a little about me, my work experience and previous projects that I've worked on. It also features a contact form equipped with reCAPTCHA and form validation to reduce spam.

Machine Learning | Deep Q-Learning

Reinforcement learning agent utilizing a CNN built with PyTorch. The agent is trained on random racetrack configurations with the goal of reaching the finish line as fast as possible while staying on the track and minimizing traction loss.

Web Development

Designed and developed a dynamic website that features a seamless and responsive user experience across various devices and screen sizes. It showcases the company's team, their services, previous projects and a contact section.

Other

Automated external defibrillator simulation application written in C++ following Agile development practices. Object-oriented design principles are applied to ensure maintainability and scalability. The GUI was implemented using Qt C++ providng an intuitive and responsive user experience. The user can work their way through various possible real-life scenarios by placing the electrode pads, checking the heart rhythm, providing shocks if necessary and perform CPR.