Ananta Sinha

Enginnering Associate @Kittelson & Associates, Inc.

Baltimore, MD, US

Developed a scalable transfer learning-based computer vision model using HRNet (PyTorch) to detect roadway features by extending output layers, fine-tuning additional classes, detecting stress level of vulnerable road users and classifying them, labeling data in safety analysis; leveraged cloud computing for scalable, high-accuracy infrastructure analysis- Built a image classification model using aerial imagery and tabular data by combining transfer learning and CNN, trained from scratch; developed a pipeline to extract and preprocess road data from OpenStreetMap, utilizing Keras, PyTorch, and Azure cloud for efficient training and deployment- Analyzed pedestrian yield data to evaluate treatment effectiveness across roadway types and speeds, and developed a Logistic regression model to predict drivers yield probability using python and statistical modeling to support safety planning- Analyzed INRIX probe speed data to diagnose freeway congestion patterns, identify contributing factors, and support statewide congestion management strategies- Developed performance metrics including travel time reliability, confidence bands, and seasonal variation to guide operational planning and investment prioritization- Produced diagnostic visuals (e.g, heat maps, congestion profiles) to support data-driven decision-making for corridor-level and systemwide planning efforts.

West Bengal University of Technology, Kolkata

Bachelor's degree, Civil Engineering

Indian Institute of Technology, Guwahati

Masters, Structural Engineering

University of Georgia College of Engineering

Doctor of Philosophy - PhD, ENGINEERING