Meet Tempo

The easiest way to create new machine products and services

What is Tempo?


Tempo is a hosted platform for building end-to-end machine learning applications using our proprietary automation technology.

Tempo automates the three steps in a machine learning endeavor


Prediction Engineering

Identify the specific problem to address your business need

Automated Feature Engineering

Intelligently extract the most important predictive signals

Machine Learning

Automatically learn rules from historical data that can be applied to new data

Prediction Engineering

Hone in on the business need

Translate a business need into historical training examples for ML algorithms to learn from. 

Make ML Pervasive

Easily define new machine learning problems across your business.

Experiment for maximum impact

Reduce the time needed to iterate on different formulations of the same problem.

Forecast Future Spending Created with Sketch. Jarom April 16, 2018 $0 $33 $113 $50 Sam March 23, 2018 John Mar 12, 2018 Feb 23, 2018 Carly CUTOFF TIME FUTURE SPENDING CUSTOMER Extract Training Examples Tempo scans your historical data to find past customers to learn form Your Training Examples Using the data from before the cutoff time, we can learn from these example customer how to predict future spending Forecast window The future period of time to calculate total spending 30 Days Lead time Increasing the lead gives you more time to act on the prediction 7 Days Spending in last month Target your customer who spent at least this much $300 EXTRACT Forecast Future Customer Spending
Deep Future Syntesis Created with Sketch. Deep Feature Syntesis 1.0 MEAN(sessions.COUNT(events)) Average number of events per session for a user 7 Max(sessions.SUM(total)) 0.5 TREND(sessions.AvgTimeBetween (start_time)) Rate of change of the average time between a user’s sessions Select Primitives Recommended Features TREND AVG. TIME BETWEEN COUNT MEAN MENU SUM SUM

Automated Feature Engineering

Make data machine learning-ready

Machine learning algorithm requires the right features, or explanatory variables, to build accurate models.

Leave no stone unturned

Deep Feature Synthesis can extract 1000s of features, automatically. ​Learn more​.

Customize to your data

Write custom primitives to extend our continually growing library.

Deep Future Syntesis Created with Sketch. Deep Feature Syntesis 1.0 MEAN(sessions.COUNT(events)) Average number of events per session for a user 7 Max(sessions.SUM(total)) 0.5 TREND(sessions.AvgTimeBetween (start_time)) Rate of change of the average time between a user’s sessions Select Primitives Recommended Features TREND AVG. TIME BETWEEN COUNT MEAN MENU SUM SUM

Machine Learning

Find the best model

Automatically search numerous modeling pipelines to find the one with best performance

Validate the model

Test your model to ensure it will produce trustworthy predictions on future data.

Deploy with ease

Export modeling code or use Tempo’s hosted APIs to put machine learning to work for you

Model Search Created with Sketch. Progress What would you like to de next? Evaluate against baseline Configure ROI Parameters Select modeling algorithm Export deployment code Calculate ROI Request data scientist audit Export deployment code View top features Select Different Model Accuracy 90% 30% better than baseline $190K Features 24 Your Model VERIFIED Using the data from before the cutoff time, we can use these example customer to learn how to predict future spending Model Search

Built for Success


Enterprises are transforming themselves using machine learning. Tempo is your path to machine learning success.

Deep Feature Synthesis

State-of-the-art automated feature engineering.
Learn more

Primitives

Reusable feature engineering building blocks to power machine learning models

Actionable Models

Interpretable models that you can explain to stakeholders

Flexible deployment

Use Tempo on-premise or in the cloud

Scalable Architecture

Scale computation to large datasets

Powerful Integrations

Integrate Tempo with popular tools for visualization, data prep, and modeling

Ready to get Started?


Contact Us