HyperOptimizer
Hyperparameter optimization infrastructure for any container
Hey Indie Hackers 👋
We’re excited to introduce HyperOptimizer, a platform designed to streamline hyperparameter optimization without the need for you to build and maintain the underlying infrastructure.
The idea for HyperOptimizer came from our experiences at autotradelab. During the development of quantitative trading strategies, we found that hyperparameter optimization was a critical component of our workflow. Manual testing of different combinations was too slow, and existing solutions were often challenging to configure, limited to specific machine-learning frameworks, or required substantial infrastructure work.
To address these issues, we decided to create our own solution.
With HyperOptimizer, you simply provide a Docker container, specify the parameters you wish to explore, and define the metric you want to optimize. Our platform then takes care of scheduling, parallel execution, tracking results, and determining which configurations to test next.
While we initially focused on quantitative trading, the methodology can easily be applied to machine learning, simulations, engineering workloads, and nearly any containerized optimization problem.
We are currently in the early stages of development and are dedicated to ensuring that the onboarding and developer experiences are as straightforward as possible.
I would love to hear how others currently approach hyperparameter or parameter optimization!
Check us out at: https://hyperoptimizer.com
2 Comments
What stands out to me is that you're not really competing with optimization frameworks.
You're competing with all the engineering work surrounding optimization. Running experiments, scheduling jobs, tracking results, and managing infrastructure often becomes a bigger bottleneck than the optimization algorithm itself. If HyperOptimizer consistently removes that operational overhead, that's a much stronger value proposition than simply helping people find better parameters.
👀