Highly motivated and results-oriented professional with a strong foundation in mechanical engineering, actively transitioning into Software Development. With a Higher Diploma in Science in Computing (2022) and continuous upskilling in web application and machine learning development, I bring a unique blend of practical experience. I have hands-on experience implementing projects with PLC controllers and industrial robots, alongside building full-stack applications.

My technical expertise spans Java, Python, JavaScript, SQL, HTML, CSS, and frameworks/libraries including React, Materialize, Bootstrap, OpenCV, Flask, TensorFlow, and Spring Boot. A significant focus of my coding journey is Machine Learning and Deep Learning, reinforced by completing AI and ML courses from Stanford and IBM (2022-2024).

I am currently developing a 'Smart-Luggage' travel app, featuring camera-based item recognition, and building a mobile robot equipped with a vision system for object recognition and command execution. My computer vision projects, including those utilizing TensorFlow.js for web-based applications and YOLOv8 with ONNX for video stream object detection, are showcased on my personal websites: roboworld.pl and roboworld.react.marekaugustyn.whshost.com.

My experience centers on developing practical applications in computer vision and natural language processing. I have successfully implemented License Plate Recognition (LPR) projects, integrating seamlessly with ONVIF-compatible cameras (e.g., TAPO). Furthermore, I specialize in building solutions powered by local Large Language Models (LLMs), including Mistral and Llama, to drive text generation and summarization tasks. My portfolio includes developing Fine-Tuning and Retrieval Augmented Generation (RAG) pipelines. Notably, I've fine-tuned the Gemma 2B model using PEFT (LoRA) on custom datasets, and for RAG, I've effectively utilized Mistral 7B to deliver robust solutions.

Cheers

Marek Augustyn