cv

PhD in Biological and Agricultural Engineering

Washington State University

Agricultural Robotics, AI in Agriculture, Precision Agriculture, Deep Learning, Machine Vision

Master of Science in Electrical Engineering

Tennessee Technological University

Image processing, Multi-sensor Data Fusion

Post-Doctoral Research Associate

University of California, Davis

• MULTIMODAL TIME SERIES MODELING: Developed and deployed deep learning models (CNN-LSTM) to analyze 3,000+ hours of sensor data to recognize worker activities and productivity, achieving >95% accuracy. Deployed models using MLflow for experiment tracking and versioning.
• LARGE-SCALE DATA PIPELINE: Built an automated pipeline using deep learning and clustering to process 100M+ GPS, IMU, and weight data for yield mapping. Leveraged AWS SageMaker and SQL for processing terabyte-scale multimodal time series data.
• MULTIMODAL PREDICTIVE MODELING: Developing a yield forecasting model fusing GPS, IMU, weight sensor, geo-tagged images, and weather data.

Technical Consultant

FFRobotics Inc.

• OBJECT DETECTION FOR ROBOTICS: Developed and integrated Mask R-CNN in TensorFlow C++ for real-time object detection, 3D localization, and obstacle identification in robotic harvester.
• TOOLING & DEBUGGING: Built custom logging tools, testing scripts, and debugging utilities for hardware–software integration.
• SOFTWARE INTEGRATION: Maintained Git-based CV pipeline for continuous development and deployment.
• OUTREACH: Explained robotic and machine learning techniques and results to technical and non-technical audiences.

Graduate Research Assistant

Washington State University (Pullman, WA)

• LABEL-EFFICIENT CV: Developed weakly-supervised image-level annotation-based CV model for object (fruit and flower) counting. Leveraged Guided Backpropagation and Class Activation Mapping to interpret model decisions, reducing annotation time by >85% compared to fully-labeled approaches.
• SEMI-SUPERVISED OBJECT LOCALIZATION: Built attention-based CNN model for object detection, density estimation, and localization using point annotations.
• FULLY-SUPERVISED DETECTION & TRACKING: Built and deployed computer vision models (using PyTorch) for detection, segmentation, and multi-object tracking of agricultural elements (e.g., flowers, fruits, branches) in complex, real-world scenes with varying lighting and occlusion.
• ROBOTICS & 3D VISION: Designed ROS-based robotic system using UR5e arm, stereo vision, and 3D point cloud processing with motion planning algorithms to enable autonomous precision tasks such as blossom thinning, targeted pollination, and branch pruning.
• TEACHING AND MENTORING: Assisted the Principal Investigator in teaching a graduate-level Machine Vision course, including developing and delivering instructional material on 'Introduction to Deep Learning,' evaluating assignments and projects, onboarding interns and graduate students, and providing mentorship and research guidance.

Graduate Research Assistant

Tennessee Technological University (Cookeville, TN)

• SENSOR DATA FUSION: Built real-time sensor fusion algorithm combining 3D volumetric scans, Electromagnetic trackers, and 2D RGB endoscopic images for 3D surgical instrument localization with 2D visualization.
• MENTORING: Mentored senior undergraduate capstone projects in programming, circuit design and fabrication, CAD design, and technical report writing.

Outstanding Reviewer

American Society of Agricultural and Biological Engineers (ASABE)

Selected as one of 21 (~2.6%) outstanding reviewers out of more than 800 reviewers for ASABE Journals in 2021.

Winner: 3MT Competition

WSU CAHNRS

Winner of WSU College of Agricultural, Human, and Natural Resource Sciences (CAHNRS) 3 Minute Thesis Competition in 2022.

Winner: ASABE Annual International Meeting (AIM) Oral Presentation

ASABE

Selected as one of the winners of ITSC Student Oral Presentation Competition at ASABE Annual International Meeting in 2021.