Professional Experience
Professional journey and key accomplishments
University of California Riverside (UCR)
ACM Project Lead
Riverside, California Jan 2025 - Mar 2025
- Managed the Advanced Data Science Project at ACM@UCR, leading the Transactional Fraud Detector project team using Graph Neural Networks (GNN) on the IEEE-CIS Fraud Detection Dataset.
- Collaborated with cross-functional teams to ensure project success and data accuracy at the University of California, Riverside.
- Engineered an AI-powered SDK for AR/VR motion capture using IMU sensors, incorporating Kalman Filter for calibration, which led to a substantial 70% improvement in motion tracking accuracy.
- Architected multimodal AI pipelines, leveraging MiniCPM LLM for context-aware sensor fusion, enabling gesture classification, movement recognition, and adaptive learning for real-time AR/VR interactions.
- Custom Drivers for Embedded Systems - GPS, Bluetooth, Wi-Fi, and Camera modules on STM32 chips, enhancing data communication by 45%.
Drozone
DevOps Engineer
Bangalore, India Jun 2023 - Dec 2023
- Spearheaded the development and automation of AgriOM and Drozone, improving deployment efficiency by 80% through CI/CD pipelines, infrastructure as code (IaC), and cloud automation using Jenkins.
- Enhanced security and scalability, reducing vulnerabilities and minimizing downtime by 40%.
- Integrated Django REST API with Google Cloud Platform (GCP) to build a scalable cloud solution. Used Docker and Kubernetes for microservices orchestration and Jenkins to automate builds and deployments.
- Research under Dr. S. N. Omkar, Chief Research Scientist, leading the AgriOM project to improve satellite farm yield prediction by 60%, while enhancing weed mapping and crop health monitoring using multi-spectral UAV imagery and computer vision.
- Optimized an image processing pipeline with YOLOv5 and Mask-RCNN, reducing analysis time by 45% and improving accuracy. Implemented Non-Maximum Suppression (NMS) to resolve 20% overlap issues in farm image analysis.
- Led an Autonomous Drone Path Optimization tool with Django to determine efficient pesticide spraying routes for pomegranate trees, leveraging the Travelling Salesman Problem (TSP) algorithm and YOLOv5 for accurate crop detection.
- Introduced features including TIFF image processing, bounding box calculations for GPS extraction, and visual path display, ensuring precise drone navigation across a farm with over 600 trees.