Dan Stinga Email and Phone Number

- Led the design and implementation of advanced Graph Neural Networks (GNN) using TensorFlow and PyTorch for a state of the art recommender system, tailored specifically for wealth management services. This involved building the system from the ground up, leveraging Torch Geometric and the Deep Graph Library (DGL) for efficient handling of graph data and models, ensuring it met the complex and dynamic needs of the finance sector.- Conducted extensive research on the latest GNN architectures, utilizing Jupyter Notebooks for interactive development and analysis. This included evaluating various models with Pandas and NumPy for data manipulation and numerical computations, enhancing the accuracy and efficiency of our wealth management recommendations. Scikit-learn was employed for preprocessing data and calculating model evaluation metrics, ensuring high-quality performance.- Lead initiatives to scale GNNs for processing vast datasets, encompassing diverse financial instruments and user profiles across multiple countries, each with its own market specifics and regulatory frameworks. This involved optimizing data processing and network training techniques to handle the scale and complexity of the data efficiently, with BigQuery facilitating large-scale data analytics and TensorFlow or PyTorch ensuring high-performance model training.- Worked closely with data scientists, software engineers, and product teams to integrate the GNN-based recommender system seamlessly into the existing technology stack. MLFlow was utilized for experiment tracking and model lifecycle management, while Git/GitHub facilitated version control and collaboration among team members.- Regularly monitored, fine-tuned, and upgraded the GNN models to adapt to evolving market conditions and client needs, guaranteeing that the recommendations remained relevant, accurate, and valuable. Vertex AI was used for managing and deploying ML models on GCP, ensuring scalable and efficient model deployment.

- Developed AI algorithms for in-cabin activity monitoring using TensorFlow, PyTorch, and OpenCV, focusing on real-time analysis with advanced CNNs.- Trained machine learning models on large image/video datasets, employing Keras for rapid prototyping, along with Pandas and NumPy for data preprocessing.- Integrated AI solutions into automotive systems in collaboration with cross-functional teams, utilizing Git/GitHub for version control and teamwork.- Conducted rigorous testing and validation of AI models, leveraging MLFlow for experiment tracking and lifecycle management to ensure reliability and efficacy.- Implemented deep learning models for monitoring driver/passenger states (e.g., fatigue, attention), enhancing in-cabin safety using computer vision techniques.- Led innovation in automotive AI with a focus on vehicle safety and comfort, utilizing TensorBoard for model training visualization and fine-tuning.- Utilized GCP for scalable cloud computing/storage, and scikit-learn for model selection/evaluation, demonstrating continuous learning and technology application.