.jpeg){kind=small}
Mission
Since 2018, Gigala has been developing AI-native tools for engineering design of physical structures — combining reinforcement learning, genetic algorithms, and finite element methods to produce optimal designs where classical methods fall short.
Gigala's core focus is intelligent topology optimization: a global, gradient-free, generalizable approach capable of handling stochastic loading and practical engineering constraints. The sequential nature of reinforcement learning extends this naturally to autonomous assembly of complex structures in the physical world — with clear applications in offshore infrastructure and space construction.
Gigala consists of two modules: intelligent topology optimization, and offshore pipelay dynamics — the latter now available as a standalone platform, Ocean Intella.
People
Giorgi Tskhondia (aka Giga; aka Georgy A. Tskhondiya) is an independent researcher and PhD in mathematical modeling with a background spanning offshore structural engineering and AI. Since 2018, his work has focused on applying reinforcement learning and evolutionary algorithms to engineering design — from topology optimization of structural components to autonomous assembly for space construction.
His research interests include intelligent topology optimization, offshore pipelay dynamics, and electroplasticity, with applications in offshore infrastructure, rocket engines, analog circuits, and lunar base construction.
Resources
Access latest news and developments of the project.
Guides
Read documentation on using Gigala models.
FAQ
Find the answers for the most frequently asked questions below
How can I participate in the project?
Reach out via ResearchGate or gigatskhondia@gmail.com to discuss collaborations, consultancy, paid features, or training.
What do I get as a sponsor?
Direct input into the product roadmap — sponsors participate in discussions about future features and help shape the direction of development.
Who is the intended audience?
Engineers, researchers, students, and technical decision-makers who want to explore AI-driven engineering design beyond conventional tools.
Is this a research or commercial product?
Both. Cutting-edge research and practical engineering are not opposites — they inform each other. Core functionality remains open source; select features and services are available commercially.
Does it have a GUI?
Not yet. Currently the software includes basic visualization. A full GUI is on the roadmap.
How does Gigala differ from other tools?
Most engineering tools bolt AI on as a feature. Gigala is built around it. Reinforcement learning and evolutionary algorithms drive the design process itself — not just the interface. LLM agents are integrated directly to assist engineers in real time.