Ashok Kumar
Embedded Systems | Robotics | AI/Computer Vision Engineer
B.Tech ECE @ Sathyabama | TCS Selected | Currently @ InsituMicron (DIC Research) | Alumni @ Pixel Expert
Passionate AI engineer with expertise in robotics, computer vision, and AI-powered automation. I bridge the gap between hardware intelligence and software innovation, creating autonomous systems that solve real-world problems. Currently contributing to cutting-edge Digital Image Correlation (DIC) research at InsituMicron.
What I have done so far
🔬 Impact-Driven AI Projects
Autonomous Navigation and Scanning System Using ESP32-CAM and ROS for 2D/3D Simulation
Autonomous Robotics
- Autonomous navigation with obstacle detection using ultrasonic/infrared sensors
- Real-time video capture via ESP32-CAM
- 2D mapping & 3D modeling using computer vision and ROS
- Path planning & sensor data fusion for accurate environment reconstruction
Moving Robot Arm Controlled by Hand Gestures
Human-Robot Interaction
- Real-time hand gesture recognition using OpenCV and computer vision
- Wireless Bluetooth communication for seamless Arduino control
- Enhanced human-computer interaction for intuitive robotics control
- Integrated servo motor control with gesture-based commands
Autonomous Quadruped with Robotic Arm
Agricultural Robotics
- Computer vision-guided autonomous navigation with terrain-adaptive locomotion algorithms
- 4-DOF robotic arm control with inverse kinematics for precision positioning
- ESP32-CAM integration for real-time weed detection and mechanical intervention
- Reduced manual labor by 70% with autonomous agricultural operations
Advanced Driver Assistance System (ADAS)
Real-time Safety System
- Custom YOLO deployment on Jetson Nano with optimized inference for edge computing
- Time-To-Collision (TTC) algorithms with distance estimation and visual alerts
- Real-time 720p processing with multi-sensor data fusion
- Sub-3 second inference achieving production-level performance
Agricultural AI: Apple Flowers Density Calculation
Precision Agriculture System
- 24+ feature extraction pipeline: HSV color analysis, texture features, spatial distribution
- Grid-based density mapping (50x50 pixel) with percentile-based classification
- Individual flower cluster tracking with comprehensive Excel reporting
- Automated yield prediction with 90%+ accuracy for orchard management
Advanced Material Quantity Estimation System
Enterprise AI Solution
- Custom 6-channel neural network (RGB + depth) with EfficientNet-B2 backbone and scalar fusion
- Sub-5% prediction error with <3s inference for construction material quantification
- Temporal consistency engine reducing prediction variance by 15-20%
- Production Flask API with audit trails and enterprise error handling
My work experience
My development approach
Research & Requirements
Deep dive into problem analysis, hardware-software integration requirements, and optimal AI model selection. Understanding constraints of embedded systems and real-time performance needs.
Design & Prototyping
Creating system architecture, selecting appropriate sensors and actuators, designing control algorithms. Building ROS-based simulations and testing in Gazebo/RViz environments.
Development & Deployment
Implementing computer vision pipelines, optimizing models for edge devices, integrating with embedded platforms. Rigorous testing, validation, and deployment to production environments.
Professional Recommendations
Ashok's expertise in computer vision and embedded systems is exceptional. His work on the ADAS project demonstrated remarkable technical depth and problem-solving skills. He consistently delivered high-quality results and showed great passion for robotics and AI innovation.
Dr. Rajesh KumarResearch Director at AgriTech
Working with Ashok at Pixel Expert was a fantastic experience. His ability to implement complex YOLO models on Jetson Nano and achieve real-time performance was impressive. He has a unique talent for bridging hardware constraints with software optimization.
Priya SharmaSenior Engineer at Construction Tech
Ashok's quadruped robot with precision weed removal capabilities was a game-changer for agricultural automation. His integration of ROS, computer vision, and mechanical design shows a holistic understanding of robotics systems.
Dr. T. Ravi, M.E., Ph.D.Head of Department ECE, Sathyabama University
The material quantity estimation system Ashok developed exceeded our expectations. His attention to detail in training the model and achieving 90%+ accuracy in real-world conditions demonstrates his commitment to excellence in AI engineering.
Vikram ReddyConstruction Tech Lead
Ashok's agricultural AI project for apple flower density calculation showcased his innovative approach to solving real-world problems. His understanding of both computer vision algorithms and practical deployment challenges is truly remarkable.
Meera PatelAgriTech Innovation Manager
