The Research Engine
Atlas
Atlas autonomously collects, extracts, and embeds research papers across seven industries — running entirely on DMF's own GPU infrastructure in Bangalore. No third-party APIs. No SaaS dependencies. It survived the loss of its database and rebuilt itself from its own archive.
12,715Research papers indexed
2023Own infrastructure since
7Active clients
50,000+Medical images processed
738,000+Words of curriculum shipped
Seven industries, one engine
Technology22 GB
Robotics11 GB
Healthcare8 GB
Energy7.9 GB
Physics2.8 GB
Cybersecurity2.8 GB
Fintech1.2 GB
Total corpus55.7 GB
How it works
Collect
Autonomous agents crawl open-access repositories and ingest PDFs.
Extract
Full-text extraction and chunking via MarkItDown.
Embed
Semantic embeddings computed on DMF's own GPUs.
Search
Full-text and vector search across the entire corpus.
Explore the research
Read real publications from DMF's team and students — open access.