CV

Research Assistant

Laboratory for Applied Informatics Research

Advanced clustering and labeling for the PATTIE database using AI techniques, improving data management and retrieval.

• Enhanced searchability and user experience through refined data categorization.

Computer Vision Summer Intern

Neural Robotics Lab

Developed a novel pipeline for monocular depth estimation, enabling accurate reconstruction of human-robot walking environments from 2D images.

• Achieved 82.6% accuracy in stair height prediction and 81.4% accuracy in depth estimation.
• Collected and made a custom image dataset using Intel RealSense depth cameras.

Python Developer

Bernhardt-Walther Lab

Developed the Python version of the Mid-level Vision Toolbox to extract structural properties from images.

• Engineered and optimized computational tools to facilitate the study of mid-level representations in both biological and artificial vision systems.

Robotics Software Engineer

Robotics and Manufacturing Automation Laboratory

Developed software for Collaborative Robots, leveraging machine learning and robotics dynamics.

• Designed a manipulator proficient in gesture recognition, obstacle avoidance, and object transfer to humans.

Computational Science Researcher

Center for Computational Social Science

Analyzed politicians’ social media use, revealing a strong correlation between Twitter engagement and followers.

• Developed a model using BERT embedding and ML techniques for classification of fraudulent behavior on Play Store.

Master of Engineering

University of Toronto

Electrical and Computer Engineering

• Reinforcement Learning
• Data Science and Analytics
• Natural Language Processing
• Cloud Computing
• Financial Engineering

Bachelor of Technology (Honors)

International Institute of Information Technology, Hyderabad

Electronics and Communication Engineering

• Computer Vision
• Data Structures
• Mobile Robotics
• Cognitive Modelling
• Processor Architecture

Lyric Mood Classification with Deep Learning

Led a team comparing deep learning and traditional machine learning approaches for mood classification in song lyrics.

• Achieved an accuracy of 85.2% using a GPT model with word2vec embedding.

Autonomous Vehicle Navigation and 3D Environment Mapping

Implemented 2D-SLAM with pose-graph optimization and loop-closure constraints for accurate positioning.

• Employed advanced algorithms for motion planning and collision avoidance in autonomous navigation.

Retinal Blood Vessels Segmentation

Built a pipeline for retinal vasculature detection using a morphological hessian-based approach and region-based Otsu thresholding for segmentation and denoising.

AI-Driven Health Monitoring System on Google Cloud Platform

Engineered an end-to-end health monitoring system using AI models to analyze respiratory pulse signals and sleep patterns.

• Showcased scalable solutions on cloud platforms.