CV

Education

Masters of Science (MS), Computer Science, University of Wisconsin-Madison, Madison, WI
GPA: 4.0 | August 2023 – May 2025 (Expected)
Bachelor of Science (BS), Computer Science & Data Science, University of Wisconsin-Madison, Madison, WI
GPA: 3.72 | August 2021 – May 2023

Machine Learning Research Assistant
Morgridge Institute for Research, Madison, WI
July 2024 – Present

Leveraged DINOv2 for Latent Representation Learning: Applied DINOv2 to learn latent representations of phenotypic responses in a pooled CRISPRi dataset with multiplexed IF readout. Customized the model for 14-channel images, fine-tuned it, and released it as open source for researchers worldwide.
Collaborated with the Spatial Technology Platform Lab at the Broad Institute: Analyzed data from D10, D21, and D28 samples, identifying strong separation in UMAP plots. Assessed the model’s discriminative capability through pairwise gene-gene classification, achieving high consistency with Cell Profiler Features and demonstrating effective single-cell embeddings analysis using the above DINOv2 fine-tuned approach.

Project Lead, Research Assistant
Madison, WI
February 2023 – June 2024

Engineered the Insect Eavesdropper: Developed a new device from an idea to achieve a 96% precision for pest sound identification, detecting even the faintest sounds produced by pests present inside plants using machine learning algorithms.
Data Analysis and Interpretation: Utilized DINO models, autoencoders, and transformers in conjunction with other machine learning techniques to analyze large volumes of acoustic data. Identified intricate patterns indicative of pest activity, leading to profound insights into pest behavior and enabling highly effective pest management strategies.
Introduced BugPulse with affordable LiDAR technology: Achieved a 30% increase in accurate species identification in challenging environments, expanding the scope of reliable data collection for pest management efforts.
Enhanced Species Identification with NERF: Utilized Neural Radiance Fields (NERF) technology to enhance species identification accuracy within the BugPulse system. By integrating NERF with LiDAR technology, achieved a 30% increase in accurate species identification, enabling more precise and efficient pest monitoring in various environmental conditions.
Helped secure $350,000 USD in research grants for the project.

Research Assistant
Madison, WI
October 2021 – December 2022

Sensor Evaluation and Robustness Testing: Conducted precision assessments of VL53L3CX and VL6180X sensors using ROS, Gazebo, and a UR5 robot arm, measuring standard deviation in distance measurements at various distances. Also examined the sensors’ tolerance to oblique angles (±40 degrees), ensuring robustness in real-world scenarios.
Real-World Calibration Precision: Successfully conducted the calibration procedure in real-world experiments with VL53L3CX and VL6180X sensors. Achieved average position precision of 3.18 mm and 7.29 mm, respectively, along with orientation precision of 0.61° and 2.01°, respectively.
Quality Assessment through 3D Scene Reconstruction: Tested and confirmed the accuracy of calibrated sensor poses through 3D scene reconstruction. Achieved average residuals of less than 2mm for the VL53L3CX sensor, indicating high-quality reconstruction of previously unseen planar objects.

Antlion Pitch Competition Winner: Secured first place and $5,000 in the Antlion Pitch Competition at the National Entomological Society of America Conference on the Insect Eavesdropper project, showcasing its innovative approach and potential impact.
Wisconsin Governor’s Business Plan Contest Semi-Finalist: Advanced to the semi-final round of the annual Wisconsin Governor’s Business Plan Contest with the Insect Eavesdropper project, selected from 52 entries from 30 communities statewide.
World AgriTech Featured Showcase: Secured $7,200 to present the Insect Eavesdropper project, recognizing its potential in agricultural innovation.
Completed the Morgridge Entrepreneurial Bootcamp (MEB): Participated in a one-week intensive training program focused on technology entrepreneurship, co-sponsored by INSITE and the Weinert Center for Entrepreneurship.
Participated and Completed the Regional NSF I-Corps Program: Gained insights into entrepreneurship and commercialization strategies for technology innovations for the Insect Eavesdropper. Selected to advance to the national level.

Geometric Calibration of Single-Pixel Distance Sensors, Carter Silverman, Dev Mehrotra, Mohit Gupta, and Michael Gleicher, IEEE Robotics and Automation Letters, July 2022, DOI: [10.1109/LRA.2022.3176453](https://doi.org/10.1109/LRA.2022.3176453)
Eavesdropping on Herbivores: Using contact microphones to quantify Plant-Insect Interactions, Dev Mehrotra, Laurence Still, Vidit Agrawal, Kimberly Gibson, James D. Crall, Emily N. Bick, PrePrint, September 24, 2024, [https://doi.org/10.1101/2024.09.23.614472](https://doi.org/10.1101/2024.09.23.614472)
Implicit ground truthing: unsupervised learning bridges validation data gaps in ecology, Dev Mehrotra, In preparation, Expected 2024.
Using SPAD Sensors for Insect Detection with NERFs, Dev Mehrotra, In preparation, Expected 2024.

Alex Arovas (2024): Supervised undergraduate student in Computer Science for sensor development projects.
Rishit Malpani (2024): Supervised undergraduate student in Computer Science for OpenCV and audio/camera sensor development projects.
Nachiket Kerai (2024): Supervised undergraduate student in Computer Science for LiDAR sensor development and Neural Scene Reconstruction (NeRF) projects.
Vidit Agrawal (2023–2024): Supervised undergraduate student in Computer Science for unsupervised learning for unlabeled acoustic projects.

Rothamsted Research, UK: Focused on integrated pest management and sensor development using machine learning.
Kansas State University: Collaborated on agricultural technology projects aimed at pest monitoring and crop protection using models for higher accuracy.
Missouri Soybean Extension: Engaged in research activities related to soybean pest management and sustainable agriculture using sensors and models.
California Extension: Worked on precision agriculture projects to enhance crop monitoring and pest control techniques by automation and ML.

Programming Languages: Python, C++, Java, R, SQL
Software & Tools: Kafka, Spark, Docker, Hadoop, Flask, ROS, ReactJS, Laravel PHP, MySQL, React Native
Machine Learning Frameworks: TensorFlow, PyTorch, Scikit-Learn, Keras, OpenCV
Other: Kubernetes, CI/CD, Agile, JIRA

WISPOLITICS, 2024: Wisconsin Governor’s Business Plan Contest: From 30 communities statewide, 52 entries advance in 2024. WISPOLITICS, February 27, 2024.
SeedWorld, 2024: 60+ AgTech Pioneers to Showcase Breakthrough Innovations at World Agri-Tech in San Francisco. KSI, February 16, 2024.
Entomological Society of America, 2024: Insect Eavesdropper: Digital Monitoring of Crop Pests Via Vibrational Signals. Entomology Today, January 17, 2024.