Sourav Bhadra

Sourav Bhadra

Deep LearningComputer VisionGeospatial AIRemote SensingMLOps

About

👋👋👋 Hello! I'm an ML researcher and geospatial data scientist with 6+ years experience in turning multimodal data into outcomes that move both the business and the science. My home turf is deep-learning-based data fusion ranging from handheld cameras, ground robots, drones, to satellites with explainability built in from the start.

Across every domain I've worked in, I keep stakeholders at the center. The win isn't the shiniest architecture, it's the outcome the people on the other side actually need. That's what decides which models I build, which I walk away from, and how I measure success.

Outside of work, I'm a husband, a father to a toddler, and a lifelong learner. I love learning new things, building things, and sharing what I learn with others. I'm also a big fan of traveling, cooking, and trying new foods. I also play guitar and sing sometimes.

Currently exploring

  • Agentic AI for the coding workflows, personal knowledge management, and small automations that compound.
  • Geospatial foundation models; their uses, how to fine-tune them for specific cases, and how to bridge them with LLMs (which they don't slot into naturally). Most of my experiments live in agriculture, but the questions generalize to other domains.

Experience

2023 — Present St. Louis, MO
  • Replaced manual phenotyping at production scale by building end-to-end ML pipelines — from model development and validation through deployment and maintenance — on automated, cloud-native workflows.
  • Shaped sensor and satellite adoption decisions by prototyping proof-of-concept solutions with cutting-edge sensors, the newest satellites, and modern model architectures.
  • Partner with cross-functional stakeholders to translate research and business objectives into deployable data science solutions.
  • PyTorch
  • TensorFlow
  • Rasterio
  • GDAL
  • GeoPandas
  • BigQuery
  • AWS
  • Airflow
  • Docker
2024 — 2024 St. Louis, MO
  • Designed and taught Geospatial Analytics, a graduate-level course covering spatial statistics, image analysis, and large-scale geospatial visualization, to 20+ graduate students within the broader context of geospatial data science.
  • Full course materials and lecture notes are publicly available on the course website.
  • Python
  • GeoPandas
  • Rasterio
  • ArcGIS
  • Google Earth Engine
2019 — 2023 St. Louis, MO
  • Improved cross-environment generalization of crop-trait estimation by ~25% by building a physics-informed transfer-learning framework that couples the PROSAIL radiative-transfer model with deep neural networks.
  • Outperformed 2D CNN baselines for plot-scale soybean yield prediction by developing end-to-end 3D CNNs (ResNet/DenseNet) that consume multi-temporal UAV imagery directly — eliminating hand-crafted feature pipelines — and scaling training from a local GPU cluster to AWS SageMaker.
  • Outperformed single-sensor baselines for seed-composition estimation by architecting multimodal fusion CNNs that integrate hyperspectral, thermal, and LiDAR data.
  • Achieved sub-centimeter multi-sensor co-registration by engineering a photogrammetric calibration pipeline based on bundle block adjustment.
  • Published 4 first-author papers in top remote-sensing journals; presented at NAPPN, AGU, and AAG.
  • Python
  • PyTorch
  • TensorFlow
  • Rasterio
  • GeoPandas
  • AWS SageMaker
  • Pix4D
2017 — 2019 Carbondale, IL
  • Enabled watershed-scale connectivity analysis previously infeasible on single workstations by implementing TauDEM-based hydrologic-connectivity models on an HPC cluster to quantify the impact of anthropogenic drainage structures.
  • Replaced manual digitization across large LiDAR datasets by training a CNN to detect bridges and culverts from high-resolution DEMs.
  • Enabled batch stream-delineation at watershed scale by automating DEM-to-stream pipelines with ArcPy.
  • Python
  • ArcPy
  • TensorFlow
  • HPC
  • ArcGIS
2016 — 2017 Dhaka, Bangladesh
  • Standardized map outputs for engineering teams by automating map-making and geospatial workflows with ArcPy, replacing manual processes for embankment and floodplain projects.
  • Informed flood-management and engineering decisions by analyzing multi-temporal satellite imagery to detect riverbank change with eCognition, ERDAS Imagine, and Google Earth Engine.
  • Python
  • ArcPy
  • Google Earth Engine
  • eCognition
  • ERDAS Imagine

Education

  1. 2019 — 2023

    Ph.D. in Geoinformatics & Geospatial Analytics

    Saint Louis University · St. Louis, MO

    Dissertation

    “Informed AI for Food Insecurity: Applications of Remote Sensing, Neural Networks and Transfer Learning for Digital Agricultural Monitoring”

    Read
    • Dissertation produced 3 first-author journal articles and 1 peer-reviewed conference proceeding.
    • Invited talks at NAPPN, AGU, and AAG.
  2. 2017 — 2019

    M.Sc. in Geography & Environmental Resources

    Southern Illinois University Carbondale · Carbondale, IL

    Thesis

    “Assessing the Impacts of Anthropogenic Drainage Structures on Hydrologic Connectivity Using High-Resolution Digital Elevation Models”

    Read
    • Recognized with the David G. Arey Memorial Award for the best Master's thesis.
  3. 2011 — 2015

    Bachelor of Urban & Regional Planning

    Khulna University of Engineering & Technology · Khulna, Bangladesh

Honors & Awards

3-Minute Thesis — Runner-Up · 2022

Saint Louis University

Second place university-wide for the talk "Fighting Food Insecurity by Seeing the Unseen."

David G. Arey Memorial Award · 2020

Southern Illinois University Carbondale

Awarded for a Master's thesis demonstrating creative thinking in natural-resource and environmental problem solving.

Ben Dziegielewski Scholarship · 2018

Dept. of Geography & Environmental Resources, SIUC

Awarded for the highest GPA in the first year of the master's program.

Projects

Peak Fall Color

Live demo

Generative AI · 2025

Peak Fall Color map interface
Problem
Travelers and leaf-peepers lack timely, location-specific forecasts of when fall foliage will peak — static seasonal calendars miss the spatial and year-to-year variation.
Solution
An interactive platform that predicts peak-foliage timing across the continental US from MODIS satellite time series using deep-learning transformers, served through a React/Leaflet map with a Gemini function-calling RAG chatbot for natural-language queries.
Impact
Turns satellite phenology science into a consumer-facing product — a full-stack showcase of generative AI, retrieval-augmented reasoning, and geospatial ML in one shipped application.

Tools

ML / AI

  • Transformers
  • MODIS NDVI
  • Gemini Function Calling
  • RAG

Full Stack

  • React
  • FastAPI
  • Leaflet

Raster4ML

120+ ★ on GitHub

Open Source · 2022

Raster4ML logo
Problem
Bringing satellite and drone imagery into a machine-learning workflow means hand-coding vegetation indices, stacking bands, and extracting zonal statistics — slow, error-prone boilerplate that demands deep geospatial expertise.
Solution
A Python library that automates the entire feature-extraction pipeline: stack rasters, compute 350+ vegetation indices, and extract statistics over shapefile geometries in a few lines of code, built on GDAL, Rasterio, and GeoPandas.
Impact
Adopted across the geospatial-ML community with 120+ GitHub stars and full ReadTheDocs documentation, lowering the barrier to remote-sensing ML for researchers and practitioners.

Tools

Core

  • Python
  • NumPy
  • Pandas

Geospatial

  • GDAL
  • Rasterio
  • GeoPandas
  • Shapely

MapLapse

PyPI package

Open Source · 2022

MapLapse logo
Problem
Static maps can't show how a region changes over time — building animated geospatial timelapses from scratch takes substantial matplotlib and GeoPandas plumbing.
Solution
A Python library that turns a shapefile and a time-indexed dataset into animated choropleth or proportional-circle maps (GIF/MP4) with a single animate() call.
Impact
Packaged on PyPI with ReadTheDocs docs, giving data scientists a reusable one-liner for temporal map storytelling.

Tools

Core

  • Python
  • Matplotlib

Geospatial

  • GeoPandas
  • Shapely
  • Fiona

SustaiN

$15K USDA SARE grant

Applied Research · 2022

SustaiN logo
Problem
Over- and under-applying nitrogen costs farmers money and pollutes waterways; growers lack field-specific, in-season guidance on how much nitrogen to apply.
Solution
A decision-support web app that fuses PlanetScope satellite imagery with gridMET weather data to generate in-season nitrogen prescription maps for corn and sorghum at field scale.
Impact
Funded by a $15,000 USDA SARE grant and built with Illinois Corn Growers (ILCORN) and the Donald Danforth Plant Science Center to raise farmer profitability while cutting nitrogen loss.

Tools

Data & ML

  • PlanetScope
  • gridMET
  • scikit-learn

Geospatial / App

  • GDAL
  • Rasterio
  • GeoPandas
  • Streamlit

PROSAIL-Net

Peer-reviewed

Research · ISPRS J. P&RS, 2024

Problem
Estimating crop biophysical traits (leaf chlorophyll, leaf angle) from hyperspectral imagery with pure deep learning generalizes poorly across fields and seasons because labeled data is scarce.
Solution
A physics-informed, dual-stream neural network that embeds the PROSAIL radiative-transfer model into transfer learning, jointly estimating leaf chlorophyll and leaf angle from UAV hyperspectral images.
Impact
Improved cross-environment generalization of trait estimation by ~25%, published as first author in the ISPRS Journal of Photogrammetry & Remote Sensing.

Tools

Methods

  • Transfer Learning
  • PROSAIL RTM
  • Dual-Stream NN

Stack

  • PyTorch
  • Hyperspectral
  • UAV

Publications

15 peer-reviewed publications · 5 first-author

Google Scholar

Selected · First Author

  1. 2024 Journal
    PROSAIL-Net: A transfer learning-based dual stream neural network to estimate leaf chlorophyll and leaf angle of crops from UAV hyperspectral images

    S. Bhadra, V. Sagan, S. Sarkar, M. Braud, T. C. Mockler, A. L. Eveland

    ISPRS Journal of Photogrammetry and Remote Sensing

  2. 2023 Journal
    End-to-end 3D CNN for plot-scale soybean yield prediction using multitemporal UAV-based RGB images

    S. Bhadra, V. Sagan, J. Skobalski, F. Grignola, S. Sarkar, J. Vilbig

    Precision Agriculture

  3. 2022 Conference
    Automatic Extraction of Solar and Sensor Imaging Geometry from UAV-borne Push-broom Hyperspectral Camera

    S. Bhadra, V. Sagan, C. Nguyen, M. Braud, A. L. Eveland, T. C. Mockler

    ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences

  4. 2021 Journal
    Assessing the impacts of anthropogenic drainage structures on hydrologic connectivity using high-resolution digital elevation models

    S. Bhadra, R. Li, D. Wu, G. Wang, B. Rekabdar

    Remote Sensing Applications: Society and Environment

  5. 2020 Journal
    Quantifying Leaf Chlorophyll Concentration of Sorghum from Hyperspectral Data Using Derivative Calculus and Machine Learning

    S. Bhadra, V. Sagan, M. Maimaitijiang, M. Maimaitiyiming, M. Newcomb, N. Shakoor, T. C. Mockler

    Remote Sensing

Collaborations

  1. Spectral enhancement of PlanetScope using Sentinel-2 images to estimate soybean yield and seed composition

    S. Sarkar, V. Sagan, S. Bhadra, F. B. Fritschi

    2024 · Scientific Reports · Journal

  2. Drone-based imaging sensors, techniques, and applications in plant phenotyping for crop breeding: A comprehensive review

    B. Gano, S. Bhadra, J. M. Vilbig, N. Ahmed, V. Sagan, N. Shakoor

    2024 · The Plant Phenome Journal · Journal

  3. Hyperfidelis: A Software Toolkit to Empower Precision Agriculture with GeoAI

    V. Sagan, R. Coral, S. Bhadra, A. Haireti, O. Al Akkad, A. Giri, F. Esposito

    2024 · Remote Sensing · Journal

  4. Soybean seed composition prediction from standing crops using PlanetScope satellite imagery and machine learning

    S. Sarkar, V. Sagan, S. Bhadra, K. Rhodes, M. Pokharel, F. B. Fritschi

    2023 · ISPRS Journal of Photogrammetry and Remote Sensing · Journal

  5. UAV Multisensory Data Fusion and Multi-Task Deep Learning for High-Throughput Maize Phenotyping

    C. Nguyen, V. Sagan, S. Bhadra, S. Moose

    2023 · Sensors · Journal

  6. An integrated machine learning and remote sensing approach for monitoring forest degradation due to Rohingya refugee influx in Bangladesh

    M. Rahaman, M. M. Morshed, S. Bhadra

    2022 · Remote Sensing Applications: Society and Environment · Journal

  7. Data-Driven Artificial Intelligence for Calibration of Hyperspectral Big Data

    V. Sagan, M. Maimaitijiang, S. Paheding, S. Bhadra, et al.

    2021 · IEEE Transactions on Geoscience and Remote Sensing · Journal

  8. A Fully Automated and Fast Approach for Canopy Cover Estimation Using Super High-Resolution Remote Sensing Imagery

    M. Maimaitijiang, V. Sagan, S. Bhadra, C. Nguyen, T. C. Mockler, N. Shakoor

    2021 · ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences · Conference

  9. Field-scale crop yield prediction using multi-temporal WorldView-3 and PlanetScope satellite data and deep learning

    V. Sagan, M. Maimaitijiang, S. Bhadra, M. Maimaitiyiming, D. R. Brown, P. Sidike, F. B. Fritschi

    2021 · ISPRS Journal of Photogrammetry and Remote Sensing · Journal

  10. Early Detection of Plant Viral Disease Using Hyperspectral Imaging and Deep Learning

    C. Nguyen, V. Sagan, M. Maimaitiyiming, M. Maimaitijiang, S. Bhadra, M. T. Kwasniewski

    2021 · Sensors · Journal