Valohai - Deep Learning for Data Scientists

by Startacus Admin

Valohai - The lowdown on the startup that's providing Deep Learning for Data Scientists...

Finland-based startup Valohai has developed a machine learning platform for data scientists to run, iterate, and collaborate on deep learning models, from experiments to inference. It is a platform to allow developers and researchers to train their own computer systems using data.

The resulting apps and programs can be used to spot specific images or predict events, and much more. Valohai provides tools for scalable machine orchestration, automated pipelines, reproducible model training, transparent team collaboration, and managed prediction serving. All of which we totally understand.

The platform requires no setup infrastructure, and it integrates with any runtime already in place, and is capable of running any machine learning code your write.

It is entirely open source, so you are not restricted. You can run TensorFlow, Keras, CNTK, Caffe, Darknet, DL4J, PyTorch, MXNet, or anything from bash scripts to C-code. Again, we’re having no trouble grasping all this, why do you ask? Write, innovate, visualise, monitor, share, and scale without limit.

This month, Valohai joined Microsoft’s UK ScaleUp programme, where it will have access to sales, marketing, and technical support. Even without this, though, Valohai is already being used in Africa to scan the elephant population, pulling data from aerial imagery.

This has helped detection accuracy increase by 11%, providing valuable data for tackling the declining number of elephants. Two Hat Security in Canada is using Valohai to train its own systems to find images of sexual abuse on the Dark Web, a famously hard to reach and monitor part of the internet often used for illegal activity.

In the company’s home, Finland, they are helping software company Reaktor to automate Finnish waterway transportation, speeding up their model development sevenfold.

The potential of Valohai is clearly demonstrated with just these three examples.

Their vision is to become the company to rely on for managing deep learning projects, and if their platform helps such varied, exciting, and important developments and processes, then we see no reason for this not to be the case.

If you like the sound of this startup, you might want to read up on:

CropSafe want to make it easy to detect crop disease

Showtime Analytics - cinema data made easy

Overwatch - scientific research management made easy

 

Subscribe to our newsletter

If you would like to receive our startup themed newsletter, full of the latest startup opportunities, events, news, stories, tips and advice, then sign up here.

• #startup
• #deep learning
• #data science
• #machine learning
• #data scientists
• #valohai
• #data

Published on: 4th November 2018