Instrument Scientist at AbbVie Inc
Role : Instrument Scientist II
Principle Investigator : Jeff Pan, PhD
Location : Lake Bluff, IL, USA
From Feb 2022 to June 2023
Worked as a member of the Specialized Research in Chaotic Systems (SPaRCS) group, a skunkworks organization tasked to invent novel instrumentation to accelerate the discovery of new biologic and small molecule drugs.
Developed a novel instrument to perform vial filling in constrained environments. Obtained exceptional results helping AbbVie’s Drug Analysis groups to accelerate their research.
Developed a novel instrument to automate time-constrained complex fluid extraction from organ-on-chip devices emulating human gut in specialized biological environment. Used a dual arm robot manipulator with smart collaboration algorithm to achieve exceptional operation speeds. The device delivered very promising results helping AbbVie’s immunology research groups.
Lead a project to build a robust instrument that consolidates chemical compounds from smaller tubes into bigger tubes for drying down. The process automates a complex compound extraction process.
Developed software architecture and performed robot programming for an instrument to increase the throughput of a mass spectrometry analysis system our group has previously built.
Instrumental in introducing latest technologies in robotics to the group which dramatically eased some of the group’s projects and increased capabilities.
Developed many new robot APIs for the group and upgraded some of the existing ones.
Contributed for various other projects of the group’s members, playing the role of a robotics expert within the group.
Technologies worked with
Cavro Omni Flex
Cavro MiniPrep
Cavro Syringe Pumps
Peristaltic Pumps
UFactory xArm
Universal Robots
NI DAQ
Actuators
ROS
Robotiq Grippers
Dual Arm Robots
C#
.NET
Visual Studio
organ-on-chip
Vial Filling
Solid Extraction
Keyence Sensors
Robot Wrappers
Dual Grippers
* * *
Software Engineer at ARIS Technology LLC
Role : Software Engineer
Supervisor: Mr. Mingu Kang, CEO of ARIS Technology LLC.
Office : 241 Oswalt Ave, Batavia, IL, 60506
From October 2020 to Jan 2022
Works
Developed a ROS1 software-stack to perform collision checking for an Autonomous Robotic 3D Inspection System (ARIS). Programmed Fanuc R-30ib mini plus controller to integrate Fanuc CRX 10iA/L collaborative Robot with our system. Deployed the system at Fiat Chrysler Automobiles - Casting Plant
Developed Python driver to support universal robots UR10e manipulator for ARIS Inspection Software
Developed C# driver to support Yaskawa HC10DT collaborative robot with ARIS Inspection Software
Worked on interfacing Fanuc M20ib Robot with ARIS Inspection Software using OPC Unified Architecture for Fanuc Robot Controllers.
Integrated third-party Autonomous Scanning Solutions with ARIS Inspection Software to achieve 100% autonomy.
Developed python web server to control ROS interface from ARIS Inspection software
Developed Mixed Reality support for ARIS Inspection System. Developed part of a holographic user interface to be used with Microsoft HoloLens 2 device with third-party ROS2 backend support.
Enhanced the mixed reality system to integrate with large gantry systems and mobile base manipulators alike.
Investigated various state of the art 3D/2D augmented reality tracking techniques to further improve our system for real world and AR world automated alignment problem.
Demonstrated our mixed reality inspection solution at Defense Manufacturing Conference Dec’21 at Denver, Colorado
Instrumental in developing backend ROS controllers for the newer version of ARIS Inspection software yet to be released.
Deployed docker container with ROS2 system to support ARIS Mixed Reality Inspection System.
Technologies worked with
C++
ROS1
ROS2
Visual Studio
KEPServer
C#
Python Web Server
Unity Game Development
Visual Studio Code
Universal Robots
Yaskawa HC10DT
Gantry System
3D inspection
2D/3D AR Tracking
Robot Wrappers
Fanuc CRX
Fanuc M20iB
docker
MoveIt
Git
Linux
Python
UR10e
HoloLens2
Mixed Reality
* * *
Research Assistant at CEEE
Role : Robotics Research Assistant
Principle Investigator : Dr. Reinhard Radermacher
Location : EEHP, Engineering Bldg, University of Maryland
From May 2019 to May 2020
Project : RoCo - A Robot Air-Conditioner
Individually built a low-cost, powerful face recognition and tracking system for a Robot Air-Conditioner (RoCo).
Responsible for System design, Integration, Software and Hardware Development, and Testing.
Implemented learning-based tracker and face recognizer with 80% tracking accuracy and 91% recognition accuracy respectively.
Used PoseNet, a deep learning based pose estimator and augmented the tracking accuracy to 95%.
Attached 3D depth sensor ensures a precise positioning of air-nozzle towards the user.
Implemented the design on Raspberry Pi 4B. Reduced the overall cost by 40% compared to previous prototype.
Deep Learning engine runs on Coral Edge TPU. Performs inferences between 32-36 real-time fps
Built a desktop version of the same RoCo prototype.
Developed a thermal-camera based face tracking system for a Forklift mounted air-conditioning system.
Responsible to report to the Client and Principle Investigator on weekly basis.
Technologies worked with
Python
OpenCV
Machine Learning
Deep Learning
Raspberry Pi 4B
cameras
Distance Sensors
Coral Edge TPU
Pose Estimation
Git
Linux
Servos and DC Motors
Electrical and Electronic Circuit design
3D Printing
Thermal Cameras
* * *
Independent Researcher at GAMMA Lab
Role : Independent Researcher
Advisor: Dr. Dinesh Manocha.
Office : 5207, Brendan Iribe Center for Computer Science and Engineering
From August 2019 to May 2020
Project - Forecasting Trajectory and Behavior of Road-Agents using Spectral Clustering in Graph-LSTMs
In a group of three worked on a novel approach to perform Traffic Forecasting and Road-Agent behavior prediction for Autonomous vehicles.
Analyzed autonomous driving datasets namely Level 5 Lyft dataset, Argoverse dataset, Apolloscape trajectory dataset and NGSIM dataset. Designed the data processing procedures to work with these datasets. In process of developing a user-friendly API to work with trajectory data of these datasets.
Instrumental in implementing state of the art behavior prediction method.
Achieved a weighted average accuracy of 91.2% for behavior prediction
Our approach reduced average prediction error by more than 54% over prior algorithms
Published a paper at IEEE Robotics and Automation Letters Journal (IEEE RAL).
Our paper was also accepted for Publication in the proceeding of IEEE/RSJ International Conference of Intelligent Robots and Systems (IROS 2020)
Technologies worked with
Python
PyTorch
LSTM's
Multiple GPU's
Linux
GitHub
virtual environments
data processing
Level 5 Lyft dataset
Argo dataset
NGSIM dataset
Apolloscape dataset
Literature Survey
Technical writing
LaTeX
* * *
Client - Lloyd's Banking Group
Worked as an Analyst for Lloyds Banking Group Client.
Performed data fixes to the Database by writing efficient SQL queries based on the client’s requirement.
Investigated and analyzed errors that occur in the database and designed data fixing procedures for new errors.
Documented a 5-month long training process during the transition of the project from Wipro Ltd. to Tata Consultancy Services Ltd. and created a complete Knowledge base for the project.
Reported directly to the clients located at Leeds, UK. Received several accolades from the client