Maxim B.

I am an ML Engineer specializing in LLM and Computer Vision. My primary focus is developing, fine-tuning, and improving machine learning models, evaluating their performance, and adapting them to specific tasks and real-world requirements. My experience includes LLM fine-tuning with LoRA, RAG-based applications, and model evaluation, as well as Computer Vision solutions using CNN and YOLO. My Computer Vision projects include models trained on datasets containing 5,000–20,000 examples, achieving 87–94% accuracy on test datasets, depending on the task. I have also worked with sequential data, including human activity recognition (HAR) using LSTM networks. Beyond model development, I can independently handle supporting stages of the ML workflow, including data collection, dataset preparation and annotation, Python-based processing pipelines, experimentation, and model integration. This allows me to develop end-to-end ML solutions, from initial data preparation and model development to evaluation and integration. I use quantitative metrics and data analysis to assess model performance and guide improvements, while considering dataset licensing and usage restrictions. What can I do for you: -LLM fine-tuning -RAG pipeline development -LLM model evalution -Evalution dataset development -LLM pipeline development, integration and deployment -Computer Vision pipeline development -Object detection -YOLO fine-tuning -Image classification -Exploratory Data Analysis -Dataset collection -Image and text annotation Tools and technologies: Python, JS, SQL, LLM, RAG, LoRA, AWS, Embeddings, ChromaDB, CUDA, LLM evalution, Hugging Face, Unsloth, PEFT, Claude Code, YOLO, ONNX, PyTorch, OpenCV, TensorFlow/Keras, TensorFlow.js, React.js, API intergation, OCR, CVAT, Jupyter etc.