Who are we?

Paul Kullich and Michael Cutler met at British Sky Broadcasting in 2010 while Paul was working on BSkyB’s upcoming addressable advertising platform for television ‘Sky AdSmart’, and Michael was delivering cutting-edge and innovative projects using Apache Hadoop and it’s ecosystem of tools.

Michael introduced Hadoop into Sky in 2008 and has been at the bleeding edge of Hadoop development ever since.   Together they used Sky’s R&D Hadoop cluster to build and business modelling and simulation tool to complex advertising business simulations across Sky’s CRM, Air Times Sales and set top box viewing data at massive scale.

There was a meeting of minds and we co-founded TUMRA in March 2012 to deliver on our shared vision.

What’s in a name – Why are we called TUMRA?

Choosing a name for your new startup is always an interesting challenge.

We considered the option of making up a some crazy random name to find an unregistered .com domain – we passed, not wanting to spend the next few years of our lives hearing “How do you spell that?” every time we mentioned the name of the company.

So for inspiration we looked to ancient language Sanskrit which is the source of all Indo-European languages.  Sanskrit can be traced back to 1500BC as a written language and much earlier when it was handed down through the generations in its spoken form.

A few hours of brainstorming and researching with the Monier Williams dictionary and instantdomainsearch.com generated 3 possible names and we were able to acquire tumra.com cheaply as it appears the demand for Sanskrit domain names is pretty weak.

Anyway Tumra, or तुम्र, means “Big”, it’s got two syllables, is easy to spell and is an interesting story as we are doing stuff with ‘Data Science’ and ‘Big Data’.

TUMRA is currently in stealth mode as we develop our data processing platform, but were delighted to be asked to speak at the May London Data Science Meetup, Michael our CTO gave a talk on real time-demographics and demonstrated real-time machine learning using live Facebook and Twitter data.

The live demo went down rather well, so we’ve committed to include live demos in all our future talks!  He’s set the bar rather high, for our next talk we might talk about the automated generation of content metadata and entity extraction from live video feeds which we’re focused on in our current sprint.

What are we planning to write about on here

  • – our thoughts and ramblings on the application of cutting edge data science
  • – highlights of progress as we from building our real-time data processing platform
  • – reviews from conferences and meet-ups we attend

Getting in touch

We’d love to get your feedback, thoughts and questions.

You can get in touch via the blog, on twitter at @PaulKullich, @cotdp or @TUMRA or swing by and visit us at TUMRA HQ at Chiswick Park.