About Eignex

Eignex is building COMBO. The name is just a list of our features: it’s Multi-variate (lots of knobs), it uses Bandit Optimization (it learns as it goes), and it’s Constraint Oriented (it won’t let you turn the knobs to a setting that blows up your server).

At its core, we’re building it because optimization is objectively cool and production infrastructure is usually a mess of “good enough” guesses. Whether you’re trying to squeeze more efficacy out of an LLM agent, refining prompts, or just figuring out which UI parameters actually work, COMBO helps automate those tweaks within the constraints of the real world. You can run it yourself on k8s or let us handle it at eignex.io.

To make the engine work, we ended up building a bunch of specialized, data-oriented tools in Kotlin. While things like combinatorial optimization and streaming statistics are the core foundations, we open-source most of the libraries we write along the way just because they’re useful and we enjoy shipping high-performance code.

Team

Rasmus Ros

Founder & Lead Developer

Rasmus specializes in search engines, AI agents, retrieval, and optimization. The technical foundation of Eignex is grounded in his PhD thesis, on Continuous Optimization at Lund University.

Contact

For general inquiries or collaboration: [email protected]