Bio
Hello! I’m Utkarsh, a Machine Learning engineer and sports analyst with a focus on cricket. As a polyglot programmer, I code mostly in Python with a growing interest in C, Golang, and Rust. When I’m not crunching data or writing code, you can find me following sports like cricket, hockey, MMA, and NFL, or exploring the world of movies.
Experience
Machine Learning Engineer
Omdena
May 2022 - Present
- Developed and deployed various machine learning models including Random Forest, XGBoost, and deep learning models (Yolov5, YOLT) for diverse business applications.
- Implemented computer vision deep neural networks for object detection, enhancing image processing capabilities.
- Engineered data pipelines using Python, Pandas, and NumPy to process and analyze large datasets, uncovering actionable insights.
- Utilized PyTorch, TensorFlow, and Keras to build and optimize neural network architectures.
- Established model versioning and experiment tracking practices using Git, DVC, and MLFlow.
- Created data visualizations with Matplotlib and Seaborn to effectively communicate findings to stakeholders.
- Collaborated with cross-functional teams to propose and implement AI-driven solutions for business challenges.
- Applied ensemble modeling techniques to combine multiple algorithms for improved model performance.
Sports(Cricket) Video Analyst
Mad About Sports
Intern
- Conducted in-depth analysis of batsmen and bowlers’ techniques using advanced video analytics and data visualization tools, identifying key strengths, weaknesses, and strategic matchups.
- Leveraged Python ecosystem (Pandas, NumPy, Matplotlib, Seaborn) and Tableau to process and visualize complex cricket performance data from multiple sources.
- Developed and presented actionable insights and data-driven recommendations, contributing to improved player performance and team strategy.
- Collaborated with a multidisciplinary team of analysts and coaches to produce high-quality analytical reports and strategic presentations.
- Applied machine learning techniques to predict player performance trends and identify areas for improvement in training regimens.