Good code deserves a better search.

Writing the algorithm is only part of the work. Then come the choices: which parameters to try, how to run enough experiments, and how to tell whether a promising result holds up. We’re building HyperOptimizer to make that search easier to run and understand.

Bring code you can run in a container, define the parameters you want to explore, and tell us what to measure. HyperOptimizer runs the trials and brings the results together, so you can spend more time on the next idea and less time managing jobs.

Trading backtests are one example. The same approach works for model training, simulations, data pipelines, and other workloads where different settings produce different outcomes. You decide what a useful result looks like.

HyperOptimizer is built by autotradelab UG (haftungsbeschränkt). If you have an integration question, a difficult workload, or feedback from your first experiment, get in touch.