Accelerate your Machine Learning

A simple API for fast, optimal parameter tuning. A state-of-the-art Bayesian optimization ensemble.

Watch us in action

See how SigOpt’s black-box optimization platform can help you amplify your machine learning models.

Learn More

Easy Management and Integration

Your models fly in and out of production and the parameters that drive them are often lost in configuration files. SigOpt provides a consistent interface, a parallel distributed scheduler for model tuning, and integrates with Python, Spark, TensorFlow, Neon, R, Java, and more.

Google and Netflix tune their models. Why don’t you?

Your machine learning models are critical to your business. You don’t put code into production without code reviews, so don’t put models into production without tuning their parameters.

For us, parameter optimization was always a secondary task, since we couldn’t dedicate the time or resources to it, and using SigOpt allows us to automate unlocking additional performance. They have an easy integration and useful web UI for tracking our models, but most importantly we saw performance improvements to the models optimized by SigOpt.

Ondrej Linda, PhD, Director of Data Science @ Hotwire

Our vehicle routing optimization algorithm uses 20+ hyperparameters that define how it explores the huge combinatorial space of this NP-complete problem. In a few weeks, SigOpt saved us months of a single engineers’ time to yield a comparable performance improvement. Whenever we change our algorithm it makes us these kind of savings all over again. It’s like having another engineer onboard.

Marin Saric, Optimoroute

SigOpt Machine Learning Examples

See SigOpt in action including deep learning, unsupervised learning with XGBoost, and logistic regression.