Aaron Mathew

Lead ML infrastructure development ➝ Develop an image management platform for accelerating machine learning capabilities in the agriculture space ➝ Coordinate with multiple software development teams to agree upon standardized formats for image data and metadata ➝ Coordinate with field teams and farmers to shepherd data collected by various systems back to NC State for analysis ➝ Coordinate with research project leads to identify how data should be organized, stored, and processed for analysis ➝ Review ~2.5TB of agricultural images for image quality and accuracy of labels applied by field teams ➝ Develop a machine learning model which integrates RGB and Depth data to estimate biomass of crops in fields ➝ Recruit and Interview students from the COE department at NC State ➝ Coordinate and manage ~4 team members to help in development tasks ➝ Provide team members with team building activities, professional development opportunities, actionable feedback and mentorship

➝ Bringing AI in the hands of farmers. ➝ Leading contributions to projects named PlantMap3D and AgImageRepository.

Advisor: Dr. Chris Reberg-Horton ➝ Built a working proof-of-concept for real time multi-camera biomass estimation tool called PlantMap3D.

Advisor - Maria Laura Cangiano. ➝ Experimented vision-based navigation for the BenchBot.

ML in Agriculture

Automated a stereo rectification pipeline for a dual camera setup.

Advisor - Dr. Edgar Lobaton ➝ Integrated the distance sensors and obstacle avoidance pipeline in the current Benchbot deployed at data collection sites in NC, MD, and TX. ➝ Collaborated with PhD students to design the functional prototype of the Benchbot, developed the dynamics, power analysis and recommended the hardware design.

➝ As a part of the R&D team, I mainly developed the POC of an IOT-enabled water level controller and detector for the company’s new product. ➝ I worked on circuit development, software – Python and Raspberry pi, sensor calibration, and developing the IoT framework. ➝ The project resulted in 98% water savings and 89% improvement from existing models. The project is deployed for testing.

↪Advisor: Dr. Shripad Bhatlawande ↪ Duties: ➝ Developed a product from idea to final product. ➝ Led a team of 3 students to create a smart wearable product for Visually Impaired people. ➝ Used SOLIDWORKS to design the system virtually and simulate the sensors. ➝ Implemented Extended Kalman Filter for localization of the user using 6 ultrasonic sensors. ➝ Integrated a Robust Simultaneous Localization and Mapping Algorithm (SLAM) to map the users path. ➝ Implemented an ML model to predict staircases to give real time updates. ➝ Embedded actuators to give pattern based vibrations as key method for information transfer to the user. ➝ Used a TOF camera to get depth data to map the environment using vision SLAM. ➝ The microcontroller used was a NVIDIA Jetson Nano. ↪ Accomplishments: Awarded Best Research Project, Secured Funding from Dassault Systèmes la fondation and IFM Engineering Ltd, showcased the project to researchers from @IITKharagpur and @IITBombay, Presented the paper in a conference.

ECE 786 : GPGPU Architecture ( by Huiyang Zhou ) ECE 506 : Architecture of Parallel Computers ( by Greg Byrd ) ECE 563 : Microprocessor Architecture ( by Eric Rotenberg )

Fall '21 - CSC/ECE 517 : Object Oriented Design and Development. - ECE 558 : Digital Imaging Systems - ECE 556 : Mechatronics Spring '22 - ECE 542 : Neural Networks - BME 590 : Rehabilitation Robotics - ECE 763 : Computer Vision Fall '22 - ECE 518 : Wearable Biosensors - BME 590 : Non Linear Control Design - ECE/ MBA 590 : Product Innovation Lab Spring '23 : - ECE650 : Research Internship

C/C++, Computer Vision, Data Structures, Digital Signal Processing, Embedded System, Networking, Telecommunications.