AI Research Intern

GalaxEye

Experience: 0+ yrs
Location: Bangalore, Karnataka, India
Job type: Full-time

Required skills

Python
PyTorch
rasterio
GDAL
Linux
Git
MLflow

About the role

About GalaxEye

GalaxEye is an Indian space tech startup dedicated to pioneering multi-sensor imaging satellites, founded by IIT Madras students and alumni and incubated at IIT Madras. The company is developing the Drishti Mission — the world's first satellite with multi-sensor capabilities that deliver all-time, all-weather imagery from space, even during nighttime and in challenging conditions. GalaxEye's technology enables advanced geospatial analyses for governments, defense, and industries worldwide.

Our AI team builds the deep learning tools that extract intelligence from high-resolution EO and SAR data coming off these satellites. This internship puts you at the heart of that work.

What You'll Work On

You'll work on hard, open research problems in satellite image understanding — the kind where the data doesn't exist in any public dataset and the models don't exist off the shelf. You'll be expected to read deeply, experiment rigorously, and think independently. This is a research role, not an implementation role.

What You'll Do

Conduct structured literature surveys on deep learning applied to EO and SAR imagery — summarising state-of-the-art methods, datasets, and benchmarks.

Reproduce and evaluate existing deep learning models on satellite imagery datasets, documenting performance and failure modes.

Train and fine-tune models on proprietary high-resolution satellite imagery, running rigorous experiments with proper tracking and ablations.

Adapt and extend deep learning architectures — CNNs, transformers, GANs, or diffusion models — to address challenges specific to high-resolution satellite data.

Collaborate closely with the research team in-office, present findings in regular reviews, and contribute to internal technical reports.

Requirements

Required Skills

Strong Python and PyTorch.

Solid foundations in deep learning — CNNs, transformers, GANs, or diffusion models.

Familiarity with raster geospatial data (GDAL, rasterio or equivalent).

Comfortable with Linux, Git, and GPU-based training pipelines.

Good to Have

Prior exposure to SAR data processing (SNAP or similar) and datasets such as Sentinel-1, ALOS-2 PALSAR, or SAR benchmark datasets (BigEarthNet, SEN12MS).

Familiarity with EO/multispectral datasets such as Sentinel-2, Landsat, or DL benchmarks like SpaceNet, BigEarthNet.

Published work or course projects in segmentation, image translation, or geospatial ML.

Experiment tracking tools (MLflow, Weights & Biases).

Benefits

What You'll Gain

Access to proprietary high-resolution SAR and EO satellite data that is not available to academic researchers anywhere — a rare opportunity even by global standards.

Direct mentorship from researchers and engineers with backgrounds from IITs and top research institutions.

Exposure to the full research lifecycle — from problem formulation and data pipelines to model training, evaluation, and reporting.

Strong interns will have the opportunity to co-author research publications and present work internally and externally.

Exceptional interns may be considered for full-time research roles at GalaxEye.

About GalaxEye

Building multi-sensor satellites for all-weather Earth observation imaging.

This page is fully interactive when JavaScript is enabled. Please enable JavaScript to apply or browse related roles.