Dmitrijs Milajevs

Dima enjoys programming since he was a teenager. He is interested in natural language processing.

He spends working days in his office surrounded by monitors and pile of scientific papers. On a weekend, he escapes the office and spends most of the day in a coffee shop somewhere in Central London. To compensate time spent sitting, he does Modern Pentathlon.

You can find him (re)tweeting as @dimazest and showing off his professional achievements on Linkedin.

Skills and interests

programming:Python
area:web development, natural language processing
storage:PostgreSQL, Pandas
vcs:mercurial, git

Who are you and what are you doing?

We are James and Uche. Together we founded Yapnak; the lunch app that brings students exclusive lunch deals for £5 and no hassle. Funnily enough, we're not tech guys (background wise) but instead both final year Queen Mary medical students. However ...

The EECS society will be making its 3rd annual visit on the 29th of November to Bletchley park (home of the code breakers) and the national museum of computing, where the first programmable computer (Colossus) was renovated.

Bletchley park is also a war museum and a part of British heritage ...

On Oct 4th, the ACM UK/IE Programming Contest, part of the ACM ICPC, will have its second distributed annual event in several locations across the UK and Ireland. This year sponsored by Google, Paypal & EF.

The contest (which runs from 10 a.m. to 3p.m.) is a UK ...

In the previous post I pointed out some difficulties of developing a reproducible experiment. Over a weekend, I've applied some of that ideas to my paper and would like to share my experience.

This is the second attempt of making this work reproducible. The first was done by Alexander Konovalov, Devasena Inupakutika, Steve Crouch and me at the Collaborations Workshop. Our aim was to reproduce an experiment and run it in Microsoft Azure.

By that time I'd already had a Vagrant box that contained all the necessary software. However, Fedora 20, the OS I used, is not supported by Azure, so we had to use another OS.

Finally, we managed to create a VM image and run the experiment in the cloud. However, deployment took most of the time, because we had to install the system dependencies manually. It was a messy list of development libraries and Python 3.3.

The Pydata conference has started and the venue has a great view to East London. If you look carefully, you can spot Queen Mary.

East London