Akashkumar Parmar

Infusing heartbeats into robotics:
Bridging mechanical designs with machine intelligence.

About Me

Hello! I'm Akashkumar Parmar, and my journey into robotics began at the University of Maryland, where I earned my Master's degree in Robotics. My academic background provided a strong foundation in mechanical design, automation, deep learning and computer vision.

Through various projects and experiences, I've gained expertise in C++, Python, ROS, SolidWorks, and machine vision. I've developed skills in designing and implementing complex robotic systems, integrating hardware and software, and optimizing automation workflows. My work includes creating autonomous systems, performing real-time object recognition, and enhancing industrial automation processes. Besides, I am constantly learning something new, however I haven't learned how to unlearn yet!

From the garage door to the Galaxy's core, I see robotics as a pivotal technology that will transform numerous sectors. This drives my passion for creating systems that improve efficiency, accuracy, and safety in the real world. I'm particularly interested in industrial automation, autonomous systems, and robotic vision, where I believe robotics can make a significant impact. My goal is to continue pushing the boundaries of what's possible with robotics, constantly learning and innovating.

Thank you for visiting my portfolio. Let's connect and drive innovation together!

Bio-Imaging and Machine Vision Lab

Research Assistant

Aug 2023 - May 2024

Directed initiatives for system-wide communication through ROS, employing Python and C++ for software development. Integrated PLC for enhanced calibration and synchronization of the conveyor, camera, pneumatic system and robot, successfully achieving precise mushroom pick-and-place operations on a running conveyor, keeping track of multiple boxes at the same time, achieving a 70% reduction in manual process.

Bio-Imaging and Machine Vision Lab

Student Intern

May 2023 - Aug 2024

Spearheaded the comprehensive 3D mechanical design and detailed 2D drafting of an advanced robotic system tailored for the mushroom packaging industry, utilizing SolidWorks. This involved meticulous planning of robotic arm placements, motor configurations, camera mounting solutions, and conveyor layouts.

Oracle Financial Software Services

Associate Consultant

Nov 2021 - Jun 2022

Innovated and automated a cloud-based portfolio generation platform on Oracle Cloud Infrastructure (OCI), exhibiting expertise in robust cloud automation workflows, utilizing Python scripting and SQL.

Larsen and Toubro Ltd.

Student Intern

Jan 2021 - May 2021

Steered formulation of a Python-based automation solution for designing and analyzing pressure vessels, lowering manual processing time from 50-60 hours to just 15-20 minutes, representing a staggering time reduction of 99.5%.

Projects

AudioVision Assist: Enabling accessibility through deep learning

An advanced deep learning model, which processes live video feeds to engender and vocally express descriptive captions, improving real-time accessibility for visually impaired individuals with audible, detailed descriptions of surroundings achieving a remarkable accuracy of 92%.




Hardware - Autonomous pick and place Robot

An autonomous robot equipped with IMU, Arduino camera, and ultrasonic sensors for advanced real-time object recognition and obstacle avoidance, illustrating a seamless integration of hardware for world interaction and navigation.


Test Driven Software Development using C++

A test-driven pair programming showcasing skills in ROS2 and C++ and simulating in Gazebo while maintaining version control using Github, employing CI/CD pipelines, agile industrial practices and comprehensive documentation.





Single view to 3D rendering

An extensive evaluation of 3D models using voxel loss, point cloud loss, and smoothing loss, enabling a detailed comparison and performance enhancement through hyperparameter tuning.


A-Star Path planning implementation

A project demonstrating the A* search algorithm with 2D visualization in matplotlib, 3D visualization in Gazebo, and real-world implementation on a TurtleBot3 robot using ROS.






PointNet-Classification-and-Segmentation-Model

A PointNet based architecture for classification and segmentation with point clouds, implementing training, testing and analysis of the model robustness.

Skills

Programming:

Python, C++, C, MATLAB, Object Oriented Programming (OOP), Data Structures, Algorithms

Design and Simulations:

ROS, RViz, Gazebo, MoveIt, SolidWorks, Ansys, AutoCAD, Creo

Libraries and Tools:

GitHub, PyTorch, Numpy, OpenCV, Pandas, Robot Studio, Doxygen, LaTeX, MS Office, Windows, Linux

Soft Skills:

Leadership, Communication, Problem Solving, Teamwork, Drive, and Initiative



Get in touch