Education
MSc in Geography (GIS Specialization) - 4.0/4.0 GPA
South Dakota State University, Brookings, SD | 2024 - 2026 | Graduate Research Assistant
Core Areas: Machine Learning & Deep Learning for Remote Sensing, High Performance Computing, Agricultural & UAS Remote Sensing, Quantitative Remote Sensing, Enterprise Data Modeling
B.E. in Geomatics Engineering
Institute of Engineering, Tribhuvan University, Pokhara, Nepal | 2016–2020
Thesis: Automated Road Extraction from Sentinel-2 Imagery using CNN
Core Areas: GIS, Remote Sensing, Photogrammetry, Digital Image Processing, Radar & LiDAR, GPS, Survey Techniques, Python for Geospatial Applications
Work Experience
Graduate Research Assistant - Remote Sensing & Digital Agriculture
I build production-grade wheat phenotyping platform utilizing computer vision methods, turning UAV, close-range, and satellite imagery into traits useful in breeding pipelines.
- Built and deployed a cloud-native phenotyping platform WheatAI.net , converting multi-source imagery into quantitative phenotypes such as wheat spike, spikelet, and kernel counts.
- Deployed satellite-based wheat tiller density mapping using Harmonized Landsat Sentinel via the GEE Python API, benchmarked against field measurements for the NCR-SARE Wheat-TDM project.
- Developed a sub-millimeter precision wheat kernel morphology extraction pipeline (length, width, area) using computer vision with ArUco fiducial-based geometric calibration, enabling high-throughput quantitative genetics selection.
- Engineered end-to-end field pipelines for Fusarium Head Blight (FHB) and Fusarium Damaged Kernel (FDK) disease severity assessment on close-range imagery.
- Designed and trained deep learning models for automated stomatal trait analysis from microscope imagery, producing scalable stomatal and pore density and morphology summaries to accelerate phenotyping and reduce manual measurement effort.
- Led UAV flight operations and field data acquisition, overseeing mission planning, sensor configuration (RGB & multispectral), and end-to-end photogrammetric processing. Produced high-resolution orthomosaics, DSM/DTM surface models, and extracted agronomic indicators including crop vigor indices, nutrient stress proxies, and yield-related metrics to support data-driven decision making.
Tech stack: PyTorch · Object Detection (YOLO-OBB, RT-DETR, SAM) · HLS · GCP · Docker · UAV phenotyping
Research Associate - Remote Sensing Analyst
Led large-scale Earth observation analytics and operational remote sensing workflows supporting vegetation monitoring, disaster response, and land-productivity assessment across multiple countries.
- Applied large-scale spatiotemporal analytics in Google Earth Engine (GEE) on multi-decadal satellite archives (Landsat, Sentinel, GRACE; 1975–2024) to quantify vegetation productivity, biomass dynamics, and land condition trends, with in-situ validation for operational decision support.
- Built automated, production-grade remote sensing pipelines for SAR-based surface water change and </b>flood mapping (Sentinel-1, ALOS-2)</b>, reducing processing latency from ~1 week to less than 24 hours and enabling near-real-time environmental monitoring.
- Designed and maintained scalable geospatial data infrastructure integrating hundreds of heterogeneous datasets, ensuring data consistency, reproducibility, and quality control across regions and time.
- Designed and maintained PostgreSQL/PostGIS geospatial data infrastructure and ETL pipelines with clean metadata and traceability.
Tech stack: ArcGIS/QGIS · Linux · Google Earth Engine · PostgreSQL · Geoserver · Google Cloud
Internships
GIS Developer Intern
Supported operational disaster-response and environmental monitoring systems through applied machine learning and scalable geospatial workflows.
- Built and deployed a U-Net–based SAR flood-mapping workflow, delivering validated flood-extent layers into operational GIS platforms.
- Integrated model outputs into production disaster-response systems supporting Sentinel-Asia initiatives.
- Developed automated geospatial ETL and validation routines ensuring CRS consistency and reproducible outputs.
Geospatial Intern (Remote)
Applied deep learning and multi-sensor satellite analytics for environmental monitoring and investigative reporting.
- Developed U-Net segmentation workflows in Google Earth Engine using Sentinel-2 imagery for road detection and surface-change analysis.
- Conducted SAR-based detection of illegal oil dumping using rapid screening pipelines in GEE.
- Performed PlanetScope-based change analysis for environmental monitoring, producing decision-ready maps and concise social media posts.
Talks
- Thapa, S., Singh, M., Ghimire, H., Koupal, D., Kaushal, S., Halder, J., Maimaitijiang, M., Sehgal, S.K.* (2025). Advancing Wheat Disease Phenotyping through YOLOv11 and YOLOv12: Automated Detection of FHB Severity and Damaged Kernels. — 2025 CANVAS, November 9-12, 2025 , Salt Lake City, UT
- Kaushal, S., Sehgal, S.K.*, Maimaitijiang, M., Billah, M.M., Singh, M., Koupal, D., Gill, H., Thapa, S., Halder, J., Ghimire, H., Subedi, S. (2025). Integrating Genomics and High-Throughput Phenotyping with Machine Learning for Predictive Breeding in Winter Wheat. — 2025 CANVAS, November 9-12, 2025 , Salt Lake City, UT
- Maimaitijiang, M., Sehgal, S.K., Kaushal, S., Billah, M.M., Janjua, U.U.R., Subedi, S., Ghimire, H., Thapa, S., Halder, J. (2025). From Image to Insight: AI-Driven Wheat Monitoring and Yield Prediction with Multi-Scale Sensing. — 2025 CANVAS, November 9-12, 2025 , Salt Lake City, UT
- Ghimire, H., Maimaitijiang, M*., Kaushal, S., Koupal, D., Poudel, K., Thapa, S., Singh, M., Subedi, S., Janjua, U. U. R., & Sehgal, S. K*. (2025). Deep Learning-Assisted Wheat Yield Estimation Through Spike and Spikelet Counting from High-Resolution Imagery. — 56th Annual SDSU Geography Convention, April 3-4 , Brookings, SD, USA
- Ghimire, H., Maimaitijiang, M*., Kaushal, S., Koupal, D., Poudel, K., Subedi, S., Janjua, U.U. R., & Sehgal, S. K*. (2025). Deep learning-based detection and counting of wheat spikes and spikelets using high-resolution field imagery for improved yield estimation. — 2025 McFadden Symposium, March 3 , Nebraska, USA