Logical Clocks

Data may be the new oil, but refined data is the fuel for AI

Logical Clocks AB is headquartered in Stockholm, Sweden, formed from members of the Distributed Computing Group at KTH – Royal Institute of Technology and RISE SICS AB. The CEO is Assoc Prof Jim Dowling (ex MySQL AB) and the Chief Scientist is Prof Seif Haridi (Hive Streaming AB co-founder).

The company is bringing hyperscale Data Science to the Enterprise through the familiar lens of a Jupyter/TensorFlow/Conda environment, but with a backend data platform that includes the world’s most scalable filesystem, scale-out compute and GPUs, and streaming analytics.

A logical clock is a mechanism for capturing chronological and causal relationships in a distributed system.

Read more about Logical Clocks here.

Logical Clocks is launching Hopsworks, a data platform for AI

Hopsworks Enterprise is an on-premise data management and data science platform in one. Hopsworks is also a Python-first platform that combines the world’s fastest Hadoop with scale-out deep learning on GPU clusters. Develop, test, and deploy your deep learning models all in a single platform.

Logical Clocks

@@logicalclocks

2/ @hopshadoop has great features for ML hackathons: (1) share GPUs between teams - 1 project per team (2) project-based security with a restricted role where users cannot upload or download data but can still train networks in the platform, (3) dataset sharing between projects

1/ Thanks to the 40+ students who participated in the Huawei Deep Learning Hackathon on @hopshadoop using Keras/Tf/Pytorch and GPUs on a large unlabelled and small labelled images. Lots of semi-supervised learning with auto-encoders and CNNs. https://t.co/kyr86G0aYM @hopshadoop

Huawei Deep Learning hackathon on the 11-12th Dec is running on #hopsworks - multi-tenant projects, GPUs, big data, @TensorFlow - all ingredients for a great hackathon. https://t.co/kUnxDMKfsL