Umesh Reddy

AI Engineer and B.Tech Computer Science Engineering (AI) graduate with hands-on experience designing, training, and deploying deep learning and computer vision models for real-time, edge, and autonomous systems. Experienced in building end-to-end AI pipelines spanning object detection, multi-object tracking, time-series/trajectory prediction, and signal-driven decision systems, and in deploying optimized models on edge platforms (NVIDIA Jetson) under real-time latency and throughput constraints. Comfortable working across the model lifecycle – from data pipelines and training to evaluation, optimization, and production deployment – using Python, PyTorch, TensorFlow, OpenCV, LSTM/CNN architectures, ROS, and CI/CD tooling.

AI Engineer - VU Dynamics Pvt. Ltd.

(2026-03)

• Design and deploy AI/deep learning solutions at scale for real-time detection, classification, tracking, and threat assessment of drones, vehicles, and humans.
• Build and optimize neural network architectures (CNN, LSTM) for object detection and recognition on live video streams, tuning for robustness, accuracy, and inference latency.
• Designed real-time computer vision and multi-object tracking pipelines, integrating detection, state estimation, and motion guidance into a single production pipeline.
• Built an LSTM-based time-series/trajectory prediction system to learn temporal motion patterns from live sensor and video data and forecast future object paths – directly applicable to signal analytics and time-series pipelines for sensor and telemetry data.
• Developed autonomous guidance and interception algorithms combining vision-derived signals with motion-control logic, an example of agentic, reasoning-driven autonomous task execution.
• Contributed AI capabilities to a Remote Controlled Weapon System (RCWS) for automated aerial threat detection, tracking, classification, and engagement support.
• Integrated CV/AI outputs with PX4, MAVLink/MAVSDK, ROS, and Gazebo offboard-control pipelines for simulation-based development, testing, and validation of autonomous systems.
• Optimized AI inference and video pipelines for edge deployment on NVIDIA Jetson, addressing latency, frame-rate, and real-time embedded processing constraints.
• Ran extensive simulation and bench-test evaluation of detection, tracking, and trajectory-prediction models, including robustness checks prior to system-level validation.

Web Development Intern - TechOctanet Services Pvt. Ltd.

(2024-05 - 2024-06)

• Built and improved full-stack web applications with modular backend architecture, REST APIs, and database integration.
• Worked in Agile sprints and daily scrums while collaborating on client requirements, debugging, and feature development.

B.Tech - Computer Science and Engineering – AI Specialization - Amrita School of Computing, Amritapuri, Kerala (2022-10)