Resume
ML and hardware engineer with end-to-end experience across sensing platforms, robotics, computer vision, distributed training, and data infrastructure. I turn ambiguous measurement problems into validated systems by connecting hardware design, acquisition software, model development, and quantitative evaluation.
Core technical skills
- Hardware and robotics: sensor integration, device orchestration, camera and illumination control, environmental control, contact microphones, LiDAR, time-of-flight sensors, UR5, ROS, Gazebo
- AI and perception: PyTorch, DINOv2, transformers, diffusion models, VQ-VAE, self-supervised representation learning, OpenCV, NeRF, computer vision, bioacoustic classification
- Software and ML infrastructure: Python, C++, DeepSpeed ZeRO-3, multi-node GPU training, Docker, experiment-control software, telemetry and metadata logging, data ingestion, dataset curation
Engineering experience
Center for Quantitative Genetics and Genomics, Aarhus University
PhD Fellow · Aarhus, Denmark · June 2025–present
- Architected a modular long-duration test platform integrating custom hardware, illumination and moisture control, continuous video capture, timestamp-aligned metadata, and experiment-management tooling.
- Implemented a repeatable image-acquisition and data-ingestion pipeline with camera control, standardized illumination, quality control, and dataset curation; produced 5,000+ curated images across species.
- Own the full system lifecycle across hardware integration, device-control software, data schemas, computer-vision pipelines, and quantitative validation.
Caicedo Lab, Morgridge Institute for Research
Machine Learning Research Assistant · Madison, Wisconsin · July 2024–May 2025
- Trained and benchmarked DINOv2, diffusion, and hierarchical VQ-VAE architectures on 1.5 million multi-channel images, increasing gene-classification AUROC by 18% relative to an ImageNet-pretrained baseline.
- Implemented DeepSpeed ZeRO-3 across 64 A100 GPUs, reducing epoch latency from 6.4 hours to 1.7 hours (3.8× higher training throughput) and lowering compute cost by 54%.
- Developed a discrete latent representation pipeline using VQ-VAE token spaces for conditional generation of unseen perturbations and temporal transitions.
Bick Lab
Project Lead · Madison, Wisconsin · February 2023–June 2024
- Led system architecture and validation for a bioacoustic detection device integrating contact microphones, signal preprocessing, self-supervised feature extraction, and classification; achieved 96% precision on faint in-plant pest signals.
- Integrated low-cost LiDAR acquisition with NeRF-based 3D reconstruction, improving species-identification accuracy by 30% in challenging capture conditions.
- Converted prototype performance into a validated technical case that contributed to $350,000 in competitive project funding.
UW Robotics and Graphics Lab
Research Assistant · Madison, Wisconsin · October 2021–December 2022
- Developed an automated extrinsic-calibration workflow for VL53L3CX and VL6180X time-of-flight sensors using ROS, Gazebo, and UR5; achieved positional accuracy of 3.18 mm and 7.29 mm and orientation accuracy of 0.61° and 2.01°.
- Performed end-to-end pose validation through 3D reconstruction, producing sub-2 mm residual error on unseen planar targets.
Education
Aarhus University
PhD in Machine Learning and Agroecology · June 2025–August 2028 (expected)
University of Wisconsin–Madison
MS in Computer Science, GPA 3.8 · August 2023–May 2025
BS in Computer Science and Data Science, GPA 3.72 · August 2021–May 2023
Publications
- Geometric Calibration of Single-Pixel Distance Sensors. Carter Silverman, Dev Mehrotra, Mohit Gupta, and Michael Gleicher. IEEE Robotics and Automation Letters 7(3), July 2022. DOI: 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, and Emily N. Bick. Preprint, September 2024. DOI: 10.1101/2024.09.23.614472
Additional technical training
- Cajal Advanced Neuroscience Training Programme—Quantitative Approaches to Behavior and Virtual Reality (2026)
- Aarhus Comprehensive Computational Entomology Summer School, ACCESS-2025 (September–October 2025)
- CaPriC PhD School on Cyber-Physical Computing Systems (October 2025)
- Morgridge Entrepreneurial Bootcamp
- NSF I-Corps Regional Program; advanced toward the national I-Corps track
Technical leadership and mentoring
- Huangjing Qin (2025): computer-vision pipelines and dataset development for behavioral phenotyping
- Alex Arovas (2024): embedded sensing systems and experimental data acquisition
- Rishit Malpani (2024): OpenCV analysis and acoustic sensor integration
- Nachiket Kerai (2024): LiDAR sensing and NeRF workflows
- Vidit Agrawal (2023–2024): unsupervised learning for unlabeled acoustic data
Cross-institutional engineering collaborations
- Rothamsted Research, UK: integrated pest management and ML-enabled sensor development
- Kansas State University: pest-monitoring and crop-protection technology
- Missouri Soybean Extension: soybean pest management and sustainable agriculture
- California Extension: precision agriculture, automation, and ML-enabled monitoring
Recognition
- Antlion Pitch Competition: first place and $5,000 award
- Wisconsin Governor’s Business Plan Contest: semi-finalist
- World AgriTech Featured Showcase: $7,200 in support
Product communication and public engagement
- Falling Walls Competition (2024)
- World AgriTech Innovation Conference (2024)
- Entomological Society of America Conference—Antlion competition pitch, winner (2024)
In the media
- Wisconsin Governor’s Business Plan Contest: 52 entries advance—WISPOLITICS, 2024
- 60+ AgTech pioneers to showcase breakthrough innovations—SeedWorld, 2024
- Insect Eavesdropper: Digital Monitoring of Crop Pests Via Vibrational Signals—Entomology Today, 2024