01 Aug
|
Recognized
|
Bengaluru
01 Aug
Recognized
Bengaluru
We are looking for a Data Annotation Engineer with hands -on experience in annotating Earth Observation (EO) and Synthetic Aperture Radar (SAR) imagery for advanced computer vision and geospatial AI tasks. You will play a critical role in building high -quality labeled datasets for object detection, semantic segmentation, and bi -temporal change detection models.
Key Responsibilities
Annotation & Labeling
- Annotate high -resolution EO (optical, multispectral) and SAR imagery (0.5 m – 3 m resolution).
- Perform pixel -level and object -level annotations for:
- Object detection (buildings, vehicles, roads, ships, infrastructure, etc.)
- Semantic & instance segmentation
- Bi -temporal change detection / change segmentation
- Handle annotations across multiple sensor modalities (EO–EO, SAR–SAR, EO–SAR).
Geospatial & Data Handling
- Work with geo -referenced raster data (GeoTIFF, NITF, HDF5, etc.).
- Ensure spatial alignment and consistency between multi -temporal and multi -sensor datasets.
- Validate annotations against ground truth, reference layers, or auxiliary GIS data.
Quality Control
- Maintain high annotation accuracy and consistency across datasets.
- Perform peer reviews and quality audits on annotated data.
- Identify edge cases, ambiguous regions, and sensor -specific artifacts (e.g., SAR speckle, layover, shadow).
Collaboration with ML Teams
- Collaborate with ML engineers and researchers to:
- Refine labeling guidelines
- Improve class definitions and taxonomy
- Provide feedback on model errors and data gaps
- Assist in creating annotation protocols and documentation.
Requirements
Required Skills & Qualifications
Core Skills
- Robust understanding of remote sensing fundamentals, especially:
- EO imagery (RGB, multispectral)
- SAR imagery (amplitude, phase, backscatter interpretation)
- Experience with annotation tasks such as:
- Bounding boxes
- Polygons
- Pixel -wise segmentation masks
- Familiarity with change detection concepts in satellite imagery.
Tools & Technologies
- Experience using annotation tools such as:
- CVAT, Labelbox, Supervisely, QGIS, ArcGIS, or similar
- Ability to work with GIS and raster data formats:
- GeoTIFF, Shapefile, JSON, COCO, Pascal VOC
- Basic scripting skills (Python preferred) for:
- Data inspection
- Annotation validation
- Format conversion