Introduction to A, the 13th Olympian
A, the 13th Olympian is a project previously known as Astraea Intelligence.
It encompasses five parts, all research-heavy. The white papers are available in the corresponding folders in the finder view for some of the projects. Some are in stealth mode. They are all interconnected to an extent.
The mission of A, the 13th Olympian is to encourage ML research without a formal PhD from home - don't just build tools, do R&D for stuff you love.
The first project is an AutoML PINN engine as OSS:
An open-source inverse O13 PINN Engine that takes governing equations & noisy sensor data, discovers unknown physical parameters in real-time and outputs physics-guaranteed dynamics models that drop into MPC/RL/state estimation pipelines.
Commercial layer provides managed compute, deployment, and continuous adaptation.
The second is the concept of Hierarchical Federated Multi-Agent RL (HFMARL), which has been adapted to a commercial use case of Energy Fleet Coordination for Virtual Power Plants (VPPs).
O13 Energy Fleet Coordination Platform is a coordination platform built on Hierarchical Federated Multi-Agent Reinforcement Learning (HFMARL) with physics-informed neural network (PINN) integration.
The platform treats VPP coordination as what it fundamentally is: a multi-agent problem where dozens of regional aggregators must learn, in real time, how to dispatch thousands of diverse assets to collectively deliver market commitments. Rather than replacing existing VPP infrastructure, the platform provides the intelligence layer that sits between market bidding and physical dispatch: it learns optimal coordination strategies through federated reinforcement learning, enforces physical constraints through embedded electrochemical and thermal models, estimates hidden asset health through inverse physics inference and preserves data privacy.
The third is designing a purely tactile world model for soft robotic manipulation (in fluid environments).
Unlike existing visuo-tactile world models that rely on camera-based gel sensors mounted on rigid grippers, our approach will operate without any visual tactile sensing. Instead, tactile perception arises from distributed strain measurements along a tendon-driven Cosserat rod manipulator, where the deforming body itself acts as the sensor. A physics-informed neural network embedding the coupled Cosserat rod equations and slender body hydrodynamics generates dense synthetic tactile fields (spatiotemporal distributions of contact force, shear stress, and strain along the manipulator surface) decomposed into tendon actuation, hydrodynamic loading, and contact contributions.
A learned world model trained on this physics-grounded synthetic data, regularized by the governing conservation laws, predicts future tactile states from sparse real sensor readings and actuation inputs. This enables model-based control policies that plan through predicted touch in environments where vision is degraded or absent, such as underwater coral manipulation, where the robot must sense and respond to delicate contact through body deformation alone.
The fourth is a hardware soft robotics project deploying a PINN.
Lastly, the fifth is the production grade system for correcting sensor drift in a specialised robotics domain with a PINN.
Contributions to OSS are welcome. For inquiries, write to mary@astraealabs.tech