DEEPAK MAURYA
Summary
I am writing to express my interest in securing a AI and Computer Vision Engineer position.
Experienced Robotics and Computer Vision Engineer with a Master's degree from IIT Jodhpur, specializing in Robotics and Computer Vision for Advanced Manufacturing and Design. Deep expertise in AI/ML and robotics software development, and sensor integration, with strong skills in AI/ML, LLM, GAN, ROS2, Python, Tensorflow, Pytorch and path/motion planning algorithms. Proven leader in guiding teams and delivering high-quality industrial autonomous systems, with a passion for deep learning and computer vision to drive innovation in AI and robotics.
Overview
4
4
years of professional experience
Work History
Senior Software Developer
Anzo Controls Pvt. Ltd.
10.2022 - Current
- Lead a dynamic team of engineers in developing cutting-edge autonomous mobile robot and fleet management system systems for industrial applications
- Spearheaded the design and development of advanced motion planning, path-planning algorithms, multi-sensor fusion and controller selection and tuning for autonomous mobile robots in ROS2
- Architected and developed fleet management software, ensuring robust and scalable solutions for long-term deployment
- Worked on fleet manager's real-time decision-making framework, UI design, backend process handling, communication protocol selection, database management, data flow streamlining and APIs development for the external WMS system
- Integrated various perception sensors including Industrial Safety LiDAR, depth cameras, and IMUs to enhance robotic navigation and obstacle avoidance capabilities
- Coordinated with cross-functional teams to align technical deliverables with business goals, ensuring timely and successful project execution
- Engineered industrial robots tailored to customer specifications, enhancing operational efficiency
- Spearheaded project coordination, ensuring timely delivery and client satisfaction
- Managed cross-functional teams, achieving up to 30% improvement in project timelines
- Responsible for the end-to-end development of a computer vision solution for automatic sorting and defect detection of packaging carton box SKUs
- Developed and trained a single CNN model capable of sorting different packaging box SKUs and detecting defects simultaneously, optimized for various input resolutions
- Improved model performance for real-time classification and defect detection on embedded device Nvidia Jetson Xavier, ensuring seamless operation in production environments
- Designed and trained a multi-task CNN that simultaneously performs SKU classification, defect detection, and quality assessment of packaging boxes
- Conducted a comparative analysis of sorting accuracy and defect detection across different CNN architectures, presenting findings to stakeholders to demonstrate model effectiveness and limitations
- Developed image processing-based algorithms to preprocess data and enhance the accuracy of defect detection for various packaging box types
- Created class diagrams and sequence diagrams for the model's architecture ensuring clarity in design and functionality
- Integration of the sorting system system with the PLC system over Modbus-TCP
- Mentored junior engineers and interns on computer vision and robotics projects, helping to expand the team's technical skill set
- Delivered high-quality code on time by effectively managing project timelines and prioritizing tasks accordingly.
- Collaborated with cross-functional teams to integrate software components seamlessly into existing systems.
- Contributed to the architecture design of complex software systems, ensuring scalability and maintainability.
ROBOTICS SOFTWARE ENGINEER
Zebu Intelligent Systems
02.2022 - 09.2022
- Led the end-to-end development of a semantic segmentation pipeline for landing site detection from aerial images, utilizing a U-Net model for precise segmentation
- Generated and curated a large-scale dataset of aerial imagery, including data preprocessing, augmentation, and labeling, ensuring a balanced and representative dataset for training
- Designed and implemented a U-Net architecture, fine-tuned for semantic segmentation, focusing on capturing intricate details of safe landing zones
- Trained and optimized the U-Net model using a combination of techniques such as learning rate scheduling, early stopping, and hyperparameter tuning to achieve high accuracy and minimize overfitting
- Conducted extensive validation and testing of the U-Net model, using metrics such as Intersection over Union (IoU) and Dice coefficient to evaluate segmentation performance across diverse terrains
- Integrated cross-validation techniques to ensure the model's generalizability and robustness, achieving consistent results across multiple test datasets
- Developed detailed reports and visualizations of the model's performance, highlighting accuracy, false positives/negatives, and suggesting improvements for operational deployment
- Presented findings and model capabilities to stakeholders, demonstrating the reliability and precision of the model in identifying safe landing zones from complex aerial imagery
- Collaborated with cross-functional teams to deploy the trained model into a real-time environment, optimizing it for inference speed on embedded systems
JUNIOR RESEARCH FELLOW (JRF)
IIT Jodhpur - ISRO
08.2021 - 02.2022
- Implemented position-based visual servoing (PBVS) for a half-humanoid robotic arm to precisely reach target objects
- Designed singularity avoidance algorithms, ensuring smooth and reliable control of the arm in complex movements
- Integrated depth sensing and camera calibration for accurate 3D positioning, optimizing inverse kinematics and velocity Jacobian for smooth path planning
- Conducted simulations in Gazebo using ROS and real-world tests, refining the system's performance for diverse scenarios
- Collaborated with ISRO IISU teams to align the solution with project specifications and operational needs
Education
M.TECH - Robotics and Computer Vision - Advanced Manufacturing & Design
Indian Institute of Technology Jodhpur
07.2021
B. TECH - Mechanical Engineering
Faculty of Engineering & Technology, GKV
07.2018
Skills
- Computer Vision
- ML/AI
- C
- Python
- SQL
- DSA
- ROS2, Nav2
- Object-oriented programming
- Design patterns
- API integration
- Client communication
- Software architecture design
- Code reviews
- Algorithm development
Timeline
Senior Software Developer
Anzo Controls Pvt. Ltd.
10.2022 - Current
ROBOTICS SOFTWARE ENGINEER
Zebu Intelligent Systems
02.2022 - 09.2022
JUNIOR RESEARCH FELLOW (JRF)
IIT Jodhpur - ISRO
08.2021 - 02.2022
B. TECH - Mechanical Engineering
Faculty of Engineering & Technology, GKV
M.TECH - Robotics and Computer Vision - Advanced Manufacturing & Design
Indian Institute of Technology Jodhpur
Publications
- Segregation of Multiple Robots Using Model Predictive Control With Asynchronous Path Smoothing, CCTA-22, 6th IEEE Control Systems Society Conference, co-author
- A model-based optimization approach with Model Predictive Control framework with constraints that do not require knowledge of size, direction, or any potential fields for segregation. With the constraints imposed on the optimization problem that generates optimal control input for Collision avoidance. Event-triggered trajectory re-planning in case the predicted and measured trajectories of the robots have substantial variations.
Hobbies and Interests
- Traveling
- Arts
- Badminton
- Table Tennis
Publications
- Segregation of Multiple Robots Using Model Predictive Control With Asynchronous Path Smoothing, CCTA-22, 6th IEEE Control Systems Society Conference, co-author
- A model-based optimization approach with Model Predictive Control framework with constraints that do not require knowledge of size, direction, or any potential fields for segregation. With the constraints imposed on the optimization problem that generates optimal control input for Collision avoidance. Event-triggered trajectory re-planning in case the predicted and measured trajectories of the robots have substantial variations.
Similar Profiles
Blake AntzoulatosBlake Antzoulatos
Marketing Analyst at C ANZO ConsultingMarketing Analyst at C ANZO Consulting
Richard SerraoRichard Serrao
Deputy Manager – Accounts Payable at Johnson Controls India Pvt. Ltd.Deputy Manager – Accounts Payable at Johnson Controls India Pvt. Ltd.
Abhishek SatiAbhishek Sati
Senior Dispute Analyst at Johnson Controls India Pvt LtdSenior Dispute Analyst at Johnson Controls India Pvt Ltd
Aaron OzorioAaron Ozorio
Software Developer / Senior Software Developer at Enigmasoft TechnologiesSoftware Developer / Senior Software Developer at Enigmasoft Technologies
SWAPNIL JAGTAPSWAPNIL JAGTAP
SENIOR QUANTITIVE DEVELOPER / SENIOR SOFTWARE DEVELOPER at MorningstarSENIOR QUANTITIVE DEVELOPER / SENIOR SOFTWARE DEVELOPER at Morningstar