Hi, I'm Denim Patel
Senior Machine Learning Engineer
Specializing in Computer Vision, Robotics, and AI
Building intelligent systems that see, learn, and adapt
About Me
I'm a Senior Machine Learning Engineer at SimpliSafe, specializing in Computer Vision, Robotics, and Autonomous Systems. My passion lies in developing intelligent systems that can perceive, understand, and interact with the world.
I hold a Master's degree from Worcester Polytechnic Institute (WPI), where I focused on Robotics and Computer Vision. With extensive experience in deep learning, visual odometry, object detection and tracking, I've worked on cutting-edge projects ranging from autonomous navigation to real-time perception systems. My expertise spans the full ML pipeline - from research and prototyping to production deployment.
Experience
Senior Machine Learning Engineer
SimpliSafe · Full-time
Jun 2024 - Present · Boston, Massachusetts, United States
Skills: Computer Vision · Deep Learning · PyTorch · AWS
Computer Vision Software Engineer
iRobot · Full-time
Dec 2020 - Jul 2024 · 3 yrs 8 mos · Pasadena, California, United States
- Led the creation of a comprehensive ground truth system for evaluating and quantifying mobile robot localization performance, including a versatile analytics framework for regression testing and A/B testing. This offered valuable insights for performance optimization and decision-making needed to deploy features to millions of robots.
- Facilitated seamless information exchange among mobile robots present in the same environment, designed methods for sharing critical environmental data, and played a key role in enhancing situational awareness, fostering improved collaboration and coordination within a robotic ecosystem.
Skills: Visual Odometry · SLAM · Data Analysis · Python · C++
Machine Learning Intern
Samsung Research America - Think Tank Team · Internship
Aug 2020 - Dec 2020 · 5 mos
- Implemented successful 3D object detection and classification techniques customized for kitchen settings, employing advanced algorithms to analyze point cloud data. Demonstrated expertise in optimizing real-time performance for dynamic kitchen scenarios, addressing challenges like occlusions and diverse object shapes, and contributing to the advancement of computer vision technologies.
- This work was presented at CES 2020 to highlight its innovative methodology and notable impact.
Skills: 3D Object Detection · Point Cloud Processing · Deep Learning · PyTorch · Python
Education
Worcester Polytechnic Institute
Master's degree, Robotics Engineering
2018 - 2020
Nirma University
Bachelor of Engineering, Mechanical Engineering with Minor specialization in Design Engineering
2013 - 2017
Skills
Object Detection
Tracking
Visual Odometry
SLAM
CNN
R-CNN
PyTorch
TensorFlow
ROS
MoveIt
Motion Planning
Path Planning
Python
C++
CUDA
OpenCV
PCL
Awards
iRobot R&D Excellence Award
iRobot
Best Robot Design Award
ABU Robocon India
Theme: Asobi - The landing disc at Pune, India
TOYOTA Award for Unique Robot Design
ABU ROBOCON International
Theme: Robominton at Yogyakarta, Indonesia
National Winner of ROBOCON 2015
ABU Robocon India
Theme: Robominton at Pune, India
National Winner of ROBOCON 2014
ABU Robocon India
Theme: Salute to parenthood at Pune, India
Projects
Age Prediction from Image!
The aim of this project was to predict the age of a person from the image.
Visual Odometry
The aim of this project is to learn the visual Odometry pipeline. Current implementation includes feature detection, feature tracking and Pose estimation using only mono-camera.
Object Detection
The aim of the project was to detect and localization of Transparent Globe. For that we explred feature based techniques as well as SOTA DL based Faster R-CNN.
Kalman filter based object tracking
Object tracking method uses Kalman filter which has constant velocity motion model and shape detection based measurement model
Camera Calibration
The goal of this excercise was to know about the Gold-standard algorithms for camera calibration. To show the application of calibrated camera AR object is placed on a chessboard.
Motion Planning(ANA*)
The goal of this mini-project is to implement Dijkstra, A* and A* variants from scratch. Any time variant ANA* refines the path quality in each iteration as shown in the gif.
SCARA Robot
We desingned and developed SCARA Robot for custom application. 3D printing, and frameworks such as merlin and repetier were explored and integrated.
OpenCV: Face recognition
The goal of this project was to learn how to use OpenCV ready algorithms such as Haar cascade for facial recognition
Rapidly Exploring Random Trees(RRT)
RRT and family is helpful in path planning in higher dimensions especially for redundunt manipulators and multi-DOF mechanisms. Implemented RRT and its variants such as RRT* and informed RRT*. RRT* refines the path quality in each iteration as shown in the gif.
ROS MoveIt!
The aim of this project was to optimize the pick and place operation of the robotic arm based on the physical features of the objects.
Mechatronic Product Development
Developed Pan-tilt motion module, which was optimized for less inertial and rigidity and all motors are mounted on base!!