Aegean AI
Build agents that see, reason, and act
From perception to language to physical manipulation, we research and train engineering teams to build AI systems that operate in the real world.
What’s New
Updated July 2026Reasoning Manipulation Agent
A VLA model that follows spoken instructions to pick, place, and shelve objects with spatial reasoning. Demonstrated by Andrew Aquino.
3D Reconstruction
A Gaussian Splatting reconstruction of the AR4 workspace, built by scanning the environment with an iPhone on a gimbal. Generated by Ayush Ramashankar (NJIT).
Agent Harnesses
Controlling a robot in plain language. Claude Code reads the workspace, reasons about the robot’s interface, and drives it end to end.
BasketTube
AI game analyzer using computer vision and video understanding to extract player actions from basketball footage.
VisDrone Aerial Detection
Drone-view video with colour-coded bounding boxes for cars, pedestrians, and other objects, streamed from the Auraison lakehouse.
Products
Our work spans three layers, from agent infrastructure to deployed applications to frontier research.
Blog
Articles, tutorials, and insights on AI developments and research.
Read the Blog
© 2026 Aegean AI