Pursuing Computer Science undergrad (B. Tech, 2023-2027), I build with AI and build around it. All my work has ranged from agentic AI and ML-powered applications to REST APIs and open-source contribution. Yet I am equally inspired by the idea that technology truly unleashes its potential when shared. As Co-Founder of LEGION at SVIT, I have inspired others to the AI world, built the tools of an ambitious event from ground zero (whiteboard), designed and developed the event systems, and created a campus space to flourish as a learner. I flourish where the code, curiosity, and leadership meet-and I am in the process of growing into an AI Forward Deployment Engineer to connect enterprise workflows with next generation multi-agent frameworks.
Technical community member & Co founder
Legion CLUB
- Collaborated with student developers on technical initiatives, workshops, and software development projects.
- Participated in discussions on AI trends and open-source tools within the developer community.
Open source contributor
Hacktoberfest / GirlScript Summer of Code
- Contributed code and resolved issues across multiple open-source repositories during Hacktoberfest 2024.
- Enhanced open-source projects, reviewed pull requests, and practiced collaborative Git workflows through GirlScript Summer of Code.
Orbital Guardian – AI Space Debris & Collision Risk Platform
Built an AI-powered platform to track live satellites and detect space debris collision risks in Earth's orbit within 24 hours. Implemented SGP4 orbital propagation for satellite tracking and trained Isolation Forest and Random Forest models for anomaly detection and collision risk assessment. Secured 1st Runner-Up (2nd place out of 105 teams) at the Astrava 2026 Hackathon, winning ₹25,000.
AquaShare – Community Groundwater Pump Sharing App
Built a mobile-first web application for farming communities to book shared groundwater pump time slots, track water usage, and report broken pumps. Designed for 30 farmers and 5 community pumps, featuring conflict-free booking, daily water extraction limits, and real-time synchronization.
Off-Road Terrain Segmentation using DINOv2
Fine-tuned a DINOv2 ViT-B/14 backbone with a custom upsampling segmentation head using PyTorch to classify off-road scenes into 10 terrain and object classes. Used combined focal and Dice loss to address class imbalance and test-time augmentation, achieving 0.5375 mIoU and 75.5% pixel accuracy across 1,002 test images, with approximately 100 ms inference per image.
B.Tech – Computer Science Engineering
Swami Vivekananda Institute of Technology, Hyderabad
10+2
Kendriya Vidyalaya Tirumalagiri
Attendee · Oct 24, 2026