Kitchener, Ontario, Canada
Hi, I'm Michael – Director of AI at FairPlay Sports Media! I lead AI initiatives focused on cutting-edge machine learning, generative AI, and large language models. My work spans applied deep learning, real-time inference systems, and predictive modeling across imaging and time-series data. My Background With a foundation in biomedical engineering and deep research experience, I've built end-to-end ML solutions in domains ranging from medical imaging to sports analytics. I specialize in designing robust, scalable systems for AI-powered products. My Focus Areas Large Language Models (LLMs): Applied experience with Google Gemini, Vertex AI SDK, and the Google ADK framework for building and deploying custom generative applications. Deep Learning: Building and fine-tuning CNNs, RNNs, transformers, and forecasting models for real-world data. Distributed Systems: Architecting ML pipelines using Apache Spark, MapReduce, and Zookeeper. Product Engineering: Cross-functional development using Python, React, and modern cloud infrastructure to ship AI features at scale. Core Skills Languages & Tools: Python, C++, Java, React, SQL, Git ML/DL Libraries: TensorFlow, Scikit-Learn, OpenCV, HuggingFace Transformers Model Types: CNNs (1D/2D), LSTMs/GRUs, SOMs, transformer-based LLMs Data Types: Imaging, time-series, natural language, structured data Cloud & Deployment: Vertex AI, Firebase, Docker, REST APIs, Agile DevOps Research: 3x first-author journal publications, technical writing, project leadership What Drives Me I'm passionate about translating complex AI technologies into real, usable products. Known as a creative problem solver, I thrive in fast-paced teams, love learning new frameworks, and always aim to blend research-grade rigor with production-ready solutions. https://github.com/MichaelBehr Have questions, or want to chat? Message me at [email protected].
Lead and manage the FairPlay (FPAI) team in developing and deploying core predictive models and AI-driven features for major global sports including soccer, NFL, NBA, NHL, and MLB. Responsible for subject matter expertise for all AI/ML related projects and roadmap items for every working group in FairPlay.
Improving and building AI models for sport predictions. Technology Stack: Python, Tensorflow, Docker, Git
Research Associate under the leadership of Dr. Dinesh Kumbhare. Spearheaded the image processing, imaging research and predictive modelling (ML/Deep Learning) for Dr. Kumbhare's lab. My research involved: -Extensive model/algorithm development in Matlab/Python involving many types of image processing, filtering, modelling, and machine learning techniques such as support vector machines, object recognition and tracking with OpenCV, image feature extraction (SURF, SIFT etc.), regression (linear,logistic), feature-space optimization, etc. - Experience with deep learning frameworks in TensorFlow, including 2D convolutional neural networks for image segmentation/detection and long short-term memory recurrent neural networks for NLP tasks (text generation) -Experience involving ultrasound and MRI imaging, including physics, operation and handling of both imaging modalities - Analyzed time-series data including forecasting (ML/Deep learning models) for EEG/EMG time series data - Skilled at multivariate statistics including PCA, LDA, Factor analysis, MANOVA, MANCOVA etc. - Comfortable managing and analyzing large datasets, and utilizing various dimensionality reduction techniques (PCA, LASSO etc.) - Experience with Git workflow, and bash scripting (Research Pipelines) - Technical writing that includes papers, grants and abstracts - Comprehensive literature review experience of the field
Performing imaging research and biomedical engineering at the Imaging Research Center, at St. Joseph's Hospital.