About Eignex
Eignex has two values: painless and effective.
Painless because the integration works wherever you run it and the constraints come with the objective. You write down what the optimizer mustn't do, and it can't improve a number by ignoring one of them.
Effective because telling Eignex what you want to optimize is straightforward and the optimizers underneath actually beat the baselines they replace.
Everything ships Apache 2.0 on GitHub from day one. If a library is making decisions in your production system, you should be able to read the lines that decide. The docs page tracks what's out.
Rasmus Ros

Eignex is a one-person operation. I work on it nights and weekends around a day job at Theca. My background is in search, optimization, and applied ML, originating from my PhD on Continuous Optimization at Lund.
If any of this sounds interesting: [email protected].