Parshad Patel

Parshad Patel
San Francisco Bay Area
Robotics Engineer at Knightscope, Inc

I love robots !

Following my dream of practical implementation of robotics and creating long term solutions aiding
human capabilities, I joined Carnegie Mellon University and earned master’s degree in Robotic
System Development in December 2014.

I find excitement in opportunities created at the nexus of different levels and fields of study for all are
connected, and the questions of how and why intrigue me. I am an entrepreneur by passion and
have successfully led and co-founded two India based start-ups in field of industrial and defense
automation. I have a propensity to learn from my experiences and channel them to my vision of
delivering the best.

I am a vigorous self-starter. My passion revolves around (in no particular order) manipulation,
mechatronic systems and machine/computer vision. I have experience working on industrial
arms(Epson and ABB), automation peripherals, and humanoid robots serving patients in hospitals.
My competencies are ROS + MoveIt!, OpenRAVE, Matlab, Arduino, Solidworks and ladder
programming on Siemens.

I am an IoT enthusiast and love tinkering different modules and open source hardwares. My
weekends are spent experimenting electronics especially ESP8266, Xbee, LoRa and sub-Ghz
network. Oh and I love STM freebies !

I currently work at Knightscope as a Robotics Engineer. The extent of my work reaches from sensor
integration to computer vision. Deep learning has been my latest fascination and its integration into
real-time reporting by the robots has become my obsession.

Growing personally (persistently) is my respite to vagaries of life. When I am not working I am
working out. Digging the opposites on Quora. And photography drools me(no brand allegiance, love
landscape and wide-angles).

I can be reached on and parshad1710 (Skype).
Robotics Engineer at Knightscope
July 2015 – Present (2 years 6 months)
Knightscope is developing technology that will predict and prevent crime. Crime has a $1+ trillion
negative economic impact on the economy and our goal is to cut it in half. Our Crime Fighting
Autonomous Data Machines – a large-scale deployment of autonomous technology, sensors,
robotics and predictive analytics – will gather real-time on-site data and combine it with existing
large data sets as well as relevant social network feeds, allowing for a breakthrough ability to map
the future in a given environment. These self-driving robots will help society build safer, engaged
communities while reducing costs and crime. More at

At Knightscope, I am a part of Advance Research Team. My major role is end-to-end integration
of features – hardware and software. I am fluent with C++/Python and enjoy being an enthusiastic
project tinkerer. I work on ROS majorly, involve in embedded level programming of sensors and
write computer vision pipelines to solve problems of object detection.

Snippet of projects worked/currently working –

– Research and integration of analog prox sensor in hardware and software stack.
– End to end thermal camera integration.
– NVidia Tegra TX1 optimisation for pedestrian detection on videos
– Face recognition.
– Developing automated learning of objects in TensorRT, Tensorflow
– Computer vision pipeline for object detection (TensorRT, Tensorflow)
– Algorithm development for reducing the image payload to server by detecting duplicate images on
edge computers.
– Inter-computer communication using various protocol to integrate multiple detection types.

Robotics and Machine Vision Engineer at TE Connectivity
February 2015 – July 2015 (6 months)
Design and implement robotic manipulation in combination with machine vision to assemble optic
fiber connector. This would be world’s first fiber connector to be assembled autonomously.
I lead software development of the project.

– Primary work included staging and scheduling of Epson robots (6 DOF and 4 DOF on Epson
RC700 and RC90 controllers) as a response to visual feedback obtained from OpenCV and
LabView and communicate over TCP/IP with HP 3488 switches to achieve actuation of pneumatic
– Wrote TCP/IP interface layer to connect OpenCV, Epson RC+ and LabVIEW using Winsock
(winsock2.h) library.

– Developed object detection (OpenCV) based alignment algorithm (+ software) which acted as
a feedback to ThorLabs actuators(LabView) to move the assembly requiring precision upto 10
microns. The algorithm detected the contours of various parts of the assembly and then aligned
its centroids in X and Y direction. This method eliminated the parallax error in perpendicular
direction and need of height adjustment for the camera. The detection was based on RGB channel
segmentation of the obtained frame. The precision achieved was ~5 microns.

Summary of tasks:
– Vacuum pick and place of assembly components using Epson robots and HP switch boards.
– Alignment using Dinolite microscopic camera and Thorlab micro-manipulators.

Controls Intern at Hstar Technologies
June 2014 – December 2014 (7 months)
Developed software to control Hstar’s humanoid robots: RoNA Meds and RoNA Lift.
– Partially integrated MoveIt! arm planning for RoNAMeds. The goal was to pick and place objects.
– Generated collision maps from Kinect’s point-cloud data for robot’s environment and collision
avoidance using Octomap.
– Developed joint level and task-space level hardware control using simulation model on Gazebo/
RViz using MoveIt! for tele-presence/operation utility.
– Developed Inverse-Kinematic command following module in 3 DOF planar and 6DOF universal
task space.
– Added gravity compensation control to existing PID control for RoNA Meds.
– Wrote bash scripts for Linux/ROS to dynamically start/stop the ROS nodes.
– Worked with EtherCAT and CANOpen Drivers and its interaction with ROS.
– Using OpenRAVE to develop quick prototype for testing arms’ reachability and testing various IK
solvers for the the 6 DOF arm.
– Experience in working with Linux Kernel.

Work involved C++ (ROS) and Python.

Co Founder at Aris Automation
July 2012 – December 2014 (2 years 6 months)
A team of Mechanical and Mechatronics engineers with a vision to deliver automated systems for
industries and defense. We provide solutions in packaging and conveyer systems.
We have two manufacturing sites at GIDC Makarpura and Savli, Vadodara.

Co Founder at Aurora Solution
February 2011 – December 2014 (3 years 11 months)
Aurora solutions is a robotics enthusiast community with aim to to connect and bridge college
education with industry. We have conducted several workshops across Gujarat and we are
constantly experimenting and bringing in nuance from experiences of our versatile team. We
are involved in training-mentoring of students, leveraging their projects with our expertise and
collaborating their knowledge with industry.
We are also a system integrator for Indian Army, EME, Vadodara providing retro-up-gradation of
their existing system with our automated solutions.

Co-Founder at Occura System
February 2014 – September 2014 (8 months)
We are a team of researchers and developers involved in delivering cohesive cloud based systems
for doctors. It will enable them to engage their patients more effectively with our automated data
capturing-saving-accessing systems integrated with our state-of-the-art hardware services.
Our vision is to create ecosystem with “Internet of Medical-hardware”.
Our mission is make healthcare systems more responsive and autonomous.

Team: Manish Sharma, Aaron McDaniel, Parshad Patel

February 2014 – Occura Systems selected as one of Top 13 finalists by Thrill Mill

Here is link to the project. I thank Aaron McDaniel for the link.

Carnegie Mellon University
Masters of Science, Robotic System Developement, 2013 – 2014
Activities and Societies: Robotics Institute
G.H. Patel College Of Engineering and Technology
B.E Mechatronics, Mechatronics, Robotics, and Automation Engineering, 2008 – 2012
Activities and Societies: International Society of Automation, GCET Student Chapter – President and
Vice President. Student Convener – Prarambh Project Club
Parshad Patel
San Francisco Bay Area
Robotics Engineer at Knightscope, Inc


  • Updated 2 years ago

To contact this candidate email

Contact using webmail: Gmail / AOL / Yahoo / Outlook

This entry was posted in . Bookmark the permalink.