Satellite Imagery Analysis
Change detection pipelines using both SAR and optical imagery — multi-sensor fusion for robust, all-weather monitoring at scale.
Geospatial AI · Engineering Studio
Production-grade infrastructure for deploying geospatial models.
Anek Labs is a research-driven engineering studio applying artificial intelligence to spatial data — from satellite imagery to agentic systems that work across terabytes of geospatial data.
Work with us See our work01 — Focus Areas
What we build
A
B
Mobile Mapping Services
Street-level imagery pipelines end to end — from automated quality inspection with vision models to extracting insights from mobile mapping data at scale.
C
Agentic Geospatial Systems
AI agents that orchestrate geospatial workflows — accelerating knowledge transfer and automating spatial reasoning pipelines.
D
Generative AI
Video generation and image enhancement workflows — production-ready generative pipelines for visual content, built on open models.
02 — Selected Work
Projects & research
/01
SAR + Optical Change Detection
Change detection pipelines combining Synthetic Aperture Radar and optical satellite imagery for multi-use-case monitoring — land cover shifts, construction tracking, and environmental change across large geographic extents.
/02
Mobile Mapping Defect Inspection via VLMs
Vision Language Models applied to street-level and mobile mapping imagery for automated quality assurance — detecting defects and data quality issues across large imagery collections.
/03
Geospatial Agent Workflows
Agent architectures that facilitate knowledge transfer and automate spatial data pipelines — working on TB-scale geospatial data without proportional manual overhead. Open-source first, cost-aware by design.
03 — Training
Geospatial AI training & workshops
Hands-on programs for colleges, research groups, and companies that need practical fluency in remote sensing, geospatial AI, and agentic spatial workflows.
A
College Workshops
Introductory and advanced sessions on remote sensing, QGIS, GeoPython, satellite data access, and applied geospatial machine learning.
B
Company Enablement
Custom training for teams building production workflows around satellite imagery, mobile mapping data, spatial analytics, and AI-assisted operations.
C
Agentic Spatial Workflows
Training on agents and VLM/LLM workflows for data discovery, QA, map review, workflow orchestration, and geospatial knowledge transfer.
D
Workshop Formats
Guest talks, half-day workshops, weekend bootcamps, and focused team sprints built around real datasets and deployable workflows.
proof
Selected talks & workshop material
Sessions delivered across NITK, BNMIT, NMIT, NIE Mysuru, ACIWRM, FOSS4G Asia, FOSS4G, QGIS India Open Day, and Let's Talk Spatial.
04 — Technology
The stack
Python
PyTorch
PostGIS
TileDB
GDAL
Xarray
Agents / MCP
Open LLMs
05 — Principles
How we work
Cost-first infrastructure
Every architectural choice is evaluated against real cost. We prefer open-source models and self-hosted infrastructure where the tradeoff makes sense at scale.
Research to production
We don't stop at notebooks. Work is designed to run at TB-scale with reproducible, deployable pipelines from day one.
Agentic acceleration
Agents are used where they genuinely reduce toil — orchestrating workflows, transferring domain knowledge, automating repetitive analysis.
Open-source first
We contribute to and rely on open geospatial tooling. Proprietary lock-in is a liability at scale.
06 — Contact
Let's build something.
Working on satellite imagery, spatial analytics, geospatial AI infrastructure, or a workshop for your team or college? We'd like to hear about it.
[email protected]