PhD student, data scientist, developer
Ads Via The Deck
PhD student, data scientist, developer
I’m Neha Narula. I’m a PhD student at MIT in Computer Science, studying distributed, scalable databases and storage systems. I’m also interested in how we find and consume news on the web. This summer I was a data scientist at Digg (formerly News.me).
For research I use a modern PC running Ubuntu. The rest of the time I use a 13 inch MacBook Air from October 2010, which I adore with all of my being. For running experiments, I spin up servers on Amazon EC2 or use a 48 core Linux machine we have in our lab.
If I’m going to be typing a lot, I need my Kinesis Advantage. I’ve had RSI problems in the past and this keyboard, though it looks strange, has helped a lot. There was a two week frustration period when I switched, but I can type much faster now. I have made the right Control key sticky, which helps a lot with emacs.
For thinking, I usually have a full-sized Rhodia notebook and a decent pen. Nothing beats writing to get the mind flowing, but I find that I need to feel like it’s ok to mar the clean blank pages.
At home I have B&W CM5 speakers, an Anthem 225 integrated amp, a Music Hall 25.2 DAC, and a record player. I download high bitrate MP3s or FLAC and hook up my laptop or the Revo to the DAC using USB. I just got a Logitech Harmony 650 remote to replace my TV, cable, and amp remotes.
On my research machine I basically live inside Chrome, emacs, a terminal, and git. I use the common assortment of unix commands, and screen on almost every server to maintain my state. To make it easy to work anywhere I ssh into my Ubuntu box from my MacBook Air and resume where I left off. I mostly program in C, C++, and Python.
At Digg, I finally started properly programming on my Mac. I used Firefox/Chrome to maintain my work/personal gmail sessions, Hipchat to keep in touch with my team, iTerm full screen, Emacs, and iPython.
Within Python I used scipy, numpy, and matplotlib a fair amount this summer. Once you know how, it’s really easy to quickly plot data. If I need a quick webserver, I’ll use Flask or Tornado. If need a datastore I’ll use MySQL or Redis.
I use RescueTime on all my machines to get a rough idea of how much work I’m getting done each week. It can be depressing to realize how much time you spend dealing with email.
I use SizeUp to use keyboard shortcuts to move windows between monitors and position them effectively. I use Evernote to store all of my random thoughts and lists because it syncs so effortlessly. There’s no Linux version so on my Ubuntu box I use the web app in Chrome, which could be better. I also love the Evernote Webclipper Chrome extension; I can use it to easily save research ideas and things I want to buy later. Mightytext is another great Chrome extension – I use it to see and respond to text messages and missed calls on my computer while I’m working, because I usually end up leaving my phone in my bag on silent and forget to check it for 8 hours at a time.
I use Pocket as a read-it-later service, but I’ve found that I never actually go back to read what I’ve saved. It seems to be write-only.
Ideally I would use my mouse as little as possible. It’s slow and makes my hands hurt – I wish more software had keyboard shortcuts.
I wish that apps worked better together on my iPad. I like the Android way that any app can create a hook to handle sharing – I wish iOS did this too. As an example, I would rather use the Chrome app as the default to open all links.
I also want a non-crippled home media system that is clean, simple, and beautiful. The idea that hardware providers have to make deals with services that I order and pay for is ridiculous (it’s even worse when they try to charge me for the pleasure; Xbox Live I’m looking at you). If I have a Netflix account, it should just work on everything.