Jul 2024 – Apr 2025 · 9 mos · Noida, Uttar Pradesh, India · On-site

• Automated parcel delineation model using SAM,
• achieving a 70% accuracy rate, which was enhanced to
• 90% via Python scripts utilizing the Geopandas package
• for data cleaning and processing.
• Developed a high-performing crop identification model
• employing Scikit, PyTorch, CNN, and ResNet50,
• successfully trained on 1,000 images and validated
• against a dataset of 200,000 images, boosting accuracy
• from 80% to 93.4%.
• Developed Python scripts focused on crop loss
• evaluation, classification, and suitability analytics for the
• Mandaya Project.
• Significantly enriched the vector layers database by 80%
• and curtailed waste by 30% through advanced Python
• scripting techniques.
• Constructed robust APIs using Flask and FastAPI for
• vector layer data covering 28 states, organized at the
• village level with corresponding PIN codes.
• Designed and implemented interactive dashboards for
• insightful data visualization through platforms like
• Streamlit and Power BI.
• Technologies:Python•QGIS•FlaskAPI•PostgreSQL•SQL•Ma
• chineLearning•CNN•TensorFlow•PyTorch•ViT•Resnet
• 50•Heroku•GitHub•SAM Model•Git•VS Code•AWS•GIS.

Aug 2022 – Jul 2024 · 1 yr 11 mos · Delhi, India

• Streamlined the Crop Health bulletin by automating
• processes, resulting in a 60% reduction in manual labor
• by saving 1200 hours annually.
• Engaged in rigorous crop vegetation mapping across 50
• districts in two states, enhancing accuracy levels to 91%.
• Utilized machine learning algorithms to elevate yield
• prediction accuracy to 80%.
• Engineered pipelines for efficient downloading of crop
• vegetation indices (MODIS NDVI & LSWI) and IMD rainfall
• data, decreasing manual efforts by 40%.
• Developed comprehensive Python scripts for Sentinel
• data extraction, effectively integrated with AWS
• SageMaker and EC2 services targeting 33 high-risk
• districts.
• Crafted multipage dashboards using Streamlit for
• visualization of crop health and weather data,
• incorporating automated CSV reporting functionalities.
• Technologies: AWS(Sagemaker,EC2-instance,S3bucket)•
• Putty•FileZilla•Python(Automation)•SQL•VScode•Jupyter
• Streamlit•AIC Pipeline•ML- Regression Models•DL and
• AI.

Mar 2021 – Aug 2022 · 1 yr 5 mos · New Delhi, Delhi, India

• Processed drone imagery data effectively to generate
• orthophotos encompassing 10,000 villages in Uttar
• Pradesh using advanced GIS tools.
• Automated multiple AgiSoft processing steps, achieving
• an 80% reduction in manual labor.
• Generated property cards for individual households in
• ArcGIS using Python.
• Managed the UP village database efficiently while
• automating statistical reporting for UP, Haryana, and
• Himachal Pradesh.
• Technologies: Excel, Arcgis, Qgis, AgiSoft, Pix4d,
• Sitescan, Python