Peak Indicators launches world’s first fully automated machine learning platform

Tallinn ML fast-tracks AI development and enables anyone to be proficient in machine learning

Chesterfield, UK, 25 June 2020: The time required to produce powerful predictive analytics for big data has been reduced from “months to mere hours”, following the introduction of Tallinn ML, a pioneering machine learning platform built by data science and analytics company Peak Indicators.

Launched today, Tallinn ML is the world’s first machine learning platform to bring automation to every stage of the machine learning workflow, providing all of the components required to build and deploy predictive models automatically, without the need for a highly skilled data scientist.

Tallinn ML includes a unique feature-engineering engine that drastically cuts the time taken to develop new predictive algorithms by generating and testing thousands of different metrics as part of the data engineering process.

Professor Paul Clough, Head of Data Science at Peak Indicators, comments: “Until now, feature engineering was a complex, manual task that accounted for around 80% of the effort involved in delivering machine learning and predictive analytics projects. It was the last major hurdle in the end-to-end automation of data science. By automating this final piece of the jigsaw, we have reduced the time it takes to develop and deploy new models from months to mere hours, and enabled anyone to quickly become proficient in machine learning and predictive analytics on a big data scale.”

“The launch of Tallinn ML is a major step forward in the world of data science, allowing people without data-science experience to deploy highly accurate machine learning and predictive analytics quickly and at significantly lower cost. Tallinn ML delivers better results with greater efficiency and is accessible to everyone.”

The launch of Tallinn ML follows a two-year development process by Peak Indicator’s data science team in Chesterfield, Derbyshire. The platform was tested in a series of technical challenges and piloted with several international organisations:

  • During testing with a global retail and consumer goods company Tallinn ML produced a predictive model in two hours that was 18% more accurate and delivered 19 times fewer false positives than one developed over a three month period by a team of experienced data scientists.
  • The use of Tallinn ML’s automated feature engineering capability to enrich employee data at a global financial services organisation improved the accuracy of its employee churn predictions by 51%. Tallinn ML produced a series of models that could be used around the world within a week of raw data being made available.
  • Peak Indicators developed a predictive algorithm using Tallinn ML for Kaggle's Titanic Competition that outperformed more than 95% of applications submitted by rival data science teams. The feature engineering work for this project was done in less than 10 minute and required zero code.



For more information please contact:

James Taylor or Nathan Makalena

Roaring Mouse Public Relations

E: /

T: +44 (0)7796 138291 / +44 (0)7538 507274

Picture to accompany this release:

Professor Paul Clough, Head of Data Science at Peak Indicators:

About Peak Indicators

Established in 2008, Peak Indicators ( has grown to become one of the UK’s leading data science and analytics organisations. The company combines best-in-breed technologies and an expert team to drive transformational results for hundreds of financial services, retail, consumer goods, public sector and telecommunications organisations. Peak’s customers include HSBC Bank, AstraZeneca, Legal & General, Nomura, Schneider Electric and GO Outdoors. 

About Tallinn

Tallinn ( is a fully-automated machine learning platform, enabling anyone to quickly become proficient in machine learning and predictive analytics on a big data scale. Based on the Apache Spark framework, Tallinn comes with the world’s first feature engineering engine designed for machine learning data analysis; automatically improving the reliability and accuracy of predictive models.






Subscribe to our Newsletter

If you enjoyed this article why not get great insights straight to your inbox

Leave a comment