AP Engineering Allotment Analytics (EAMCET 2025)
EAMCET 2025 First Phase Counseling – Andhra Pradesh
Excited to share my project where I analyzed EAMCET 2025 first phase counseling data and built a dashboard to visualize student allotments across colleges.
Analyzed Top 5000 ranks and their college allotments
Visualized branch-wise and category-wise trends
Campus Eye
AI-powered campus monitoring system.
The goal was simple yet impactful — to analyze live video feeds, count people, and provide real-time insights through an interactive web dashboard
HR Analytics Dashboard – Workforce Insights
Designed a Tableau dashboard that transforms HR data into actionable insights, analyzing employee demographics, departmental spread, and salary trends across locations.
MotionMorph – AI-Driven Stickman Animation from Video
Developed a computer vision–based project that extracts 33 human pose keypoints from video using MediaPipe, stores them in a structured .txt file, and uses Unity (C#) to simulate and animate a stickman by mapping joints and connections from the stored pose data.
Real-Time People Counting on Escalators using YOLOv8
Built a real-time computer vision system using YOLOv8 and OpenCV to detect, track, and count people moving upward and downward on escalators, optimized for low-latency performance on resource-constrained systems.
Real-Time Vehicle Counting System Using YOLOv8
Created a computer vision system that uses YOLOv8 for object detection and OpenCV for real-time processing to count vehicles within defined road regions, providing reliable traffic insights for smart city applications.
CampusEye – AI-Powered Smart Surveillance
Collaborated in a cross-functional team to architect an end-to-end monitoring system using YOLOv8 and SORT. Engineered a low-latency AI pipeline using FastAPI and WebSockets to stream MJPEG live video and metrics, and integrated Cloudflare Tunnels for secure remote accessibility.
MotionMorph – AI Video-to-Stick Figure Pose System
Developed a computer vision system that converts real-world human motion into animated stick-figure skeletons using pose estimation. Used MediaPipe pose landmarks to extract 3D body keypoints and reconstruct motion dynamics frame-by-frame.
Logistics Flow Engine – Real-Time Escalator Passenger Tracking
Developed a real-time computer vision pipeline using YOLOv8-Nano to detect and track passengers in escalator streams. Implemented centroid-based tracking with a virtual counting line for dual-direction passenger flow analysis and a low-latency OpenCV video processing pipeline.