Resume
PROFESSIONAL EXPERIENCE - 3 YEARS SKILLS
Perception Engineer
TartanSense
April 2019 � July 2019
Bangalore, India
� Lead Computer Vision Engineer - Weed detection and Localisation:
Trained a deep learning weed detection model and localised 3D location
of weed by calibrating camera extrinsic and intrinsic parameters
� Project Manager for data collection Rover: Managed a team for
building, testing and deploying a data collection rover for collecting
image data of weeds from farms across 10 locations in India.
Sr Computer Vision Engineer
M.S. in Robotics - (3.93/4)
Soliton Technologies
May 2016 � Feb 2019
Bangalore, India
� Developer - Image Depth Categorisation using deep learning :
? Classified images into one of four categories: Close-up, Medium,
Long and Ultra long range shot.
? Generated monocular depth maps and created a four channel
RGBD image.
? Experimented with different Deep Learning Architectures and did
hyper parameter tuning to get an accuracy of 85 percent.
� Developer - Seat Belt detection:
? Constructed a sliding window detector by training an SVM classifier
on HoG features.
? Performed Hard Negative mining and Non-Maximum suppression
to get a final IoU of 75% for detector.
ACADEMIC PROJECTS
� Navigation and SLAM on a Turtlebot: Constructed a wheeled robot
navigation and EKF filter based SLAM from scratch and tested on a
turtlebot. Coded project in C++ inside RoS framework.
� Survey on Visual SLAM:
? Read 51 papers in area of visual SLAM and wrote a report style
paper by summarizing knowledge gained.
? Focused attention to distribute papers across different SLAM
frameworks and complementary sensors (inertial sensors, depth
cameras)
� Zero shot imitation: Applied a self supervised deep learning technique
to reinforcement learning problem of estimating an effective policy in
pytorch to enable a Baxter to manipulate non-rigid bodies (tying a
knot).
� Baxter, lego builder:
? Programmed a Baxter was to build a lego pyramid in RoS.
? Implemented a computer vision node for recognizing AR tag, red
lego blocks and estimating inverse projection to find 3D location
of blocks.
? Setup RoS pipeline for whole project in python.
INTERIM PROJECTS
� Object Manipulation using Youbot
in Simulation: Implemented a PI feed
forward controller for a 4 wheeled mobile
robot with a 5 DoF arm for an object
manipulation task.
� Optical Character Verification in Soliton
Smart Camera: Calculated a Euclidean
transformation to align input image with
a template image based on ORB feature
matching and verified characters present
within 150ms in ARM processor.
� ML/AI projects: Coded a UKF filter for
robot localisation , PI controller with
A star planner, Locally weighted Linear
Regression, RRT algorithms.
� Team Leader - Card Reader:
? Guided an intern to develop a
government card reader application.
? Estimated homography for aligning
card by detecting its edges in python.
? Detected text using SWT, segmented
characters and recognized each
character using DL OCR model.
EDUCATION:
Northwestern, Illinois
2019 � 2020
B.E. in Electronics & Communication
Anna University, Chennai
2012 � 2016
LEADERSHIP SKILLS
Project Coordinator, Science: Built
and managed a team of 30 volunteers
for teaching science to about 150
underprivileged kids.
Evanston
Areas: SLAM, Robotics perception,
Computer Vision, 3D vision, Machine
learning, Deep learning, Algorithms, and
Data Structures,
Languages: Python, C , C++
OS known: Linux
Tools: Vim, Bash, Git, RoS
Libraries: Pytorch, OpenCV
PROFESSIONAL EXPERIENCE - 3 YEARS SKILLS
Perception Engineer
TartanSense
April 2019 � July 2019
Bangalore, India
� Lead Computer Vision Engineer - Weed detection and Localisation:
Trained a deep learning weed detection model and localised 3D location
of weed by calibrating camera extrinsic and intrinsic parameters
� Project Manager for data collection Rover: Managed a team for
building, testing and deploying a data collection rover for collecting
image data of weeds from farms across 10 locations in India.
Sr Computer Vision Engineer
M.S. in Robotics - (3.93/4)
Soliton Technologies
May 2016 � Feb 2019
Bangalore, India
� Developer - Image Depth Categorisation using deep learning :
? Classified images into one of four categories: Close-up, Medium,
Long and Ultra long range shot.
? Generated monocular depth maps and created a four channel
RGBD image.
? Experimented with different Deep Learning Architectures and did
hyper parameter tuning to get an accuracy of 85 percent.
� Developer - Seat Belt detection:
? Constructed a sliding window detector by training an SVM classifier
on HoG features.
? Performed Hard Negative mining and Non-Maximum suppression
to get a final IoU of 75% for detector.
ACADEMIC PROJECTS
� Navigation and SLAM on a Turtlebot: Constructed a wheeled robot
navigation and EKF filter based SLAM from scratch and tested on a
turtlebot. Coded project in C++ inside RoS framework.
� Survey on Visual SLAM:
? Read 51 papers in area of visual SLAM and wrote a report style
paper by summarizing knowledge gained.
? Focused attention to distribute papers across different SLAM
frameworks and complementary sensors (inertial sensors, depth
cameras)
� Zero shot imitation: Applied a self supervised deep learning technique
to reinforcement learning problem of estimating an effective policy in
pytorch to enable a Baxter to manipulate non-rigid bodies (tying a
knot).
� Baxter, lego builder:
? Programmed a Baxter was to build a lego pyramid in RoS.
? Implemented a computer vision node for recognizing AR tag, red
lego blocks and estimating inverse projection to find 3D location
of blocks.
? Setup RoS pipeline for whole project in python.
INTERIM PROJECTS
� Object Manipulation using Youbot
in Simulation: Implemented a PI feed
forward controller for a 4 wheeled mobile
robot with a 5 DoF arm for an object
manipulation task.
� Optical Character Verification in Soliton
Smart Camera: Calculated a Euclidean
transformation to align input image with
a template image based on ORB feature
matching and verified characters present
within 150ms in ARM processor.
� ML/AI projects: Coded a UKF filter for
robot localisation , PI controller with
A star planner, Locally weighted Linear
Regression, RRT algorithms.
� Team Leader - Card Reader:
? Guided an intern to develop a
government card reader application.
? Estimated homography for aligning
card by detecting its edges in python.
? Detected text using SWT, segmented
characters and recognized each
character using DL OCR model.
EDUCATION:
Northwestern, Illinois
2019 � 2020
B.E. in Electronics & Communication
Anna University, Chennai
2012 � 2016
LEADERSHIP SKILLS
Project Coordinator, Science: Built
and managed a team of 30 volunteers
for teaching science to about 150
underprivileged kids.
Evanston
Areas: SLAM, Robotics perception,
Computer Vision, 3D vision, Machine
learning, Deep learning, Algorithms, and
Data Structures,
Languages: Python, C , C++
OS known: Linux
Tools: Vim, Bash, Git, RoS
Libraries: Pytorch, OpenCV