#top coder

[lifeatka]

A Khanversation with...analytics whiz Tony Liu

There are those people you meet in life, and you know they are just really nice. They listen patiently, they are very thoughtful, and seem like they can be anyone's friend.  Tony Liu is one of these people.

So, when I started researching for our Khanversation, I was fascinated that I knew his polite, easy-going personality, and had never heard a hint at the genius within. I might have guessed that this MIT grad is pretty great at analytics since he's one of the small handful of people on Khan Academy's analytics team trying to ensure that world can learn more effectively. However, I might not have realized that he worked for Richard Rusczyk on the Art of Problem Solving. I would never have guessed that he got a perfect score on the AIME, the qualifying exam for the USA Math Olympiad; or that he received an honorable mention at that olympiad, and twice was honorably mentioned at the Putnam (another really, really famous math competition). Oh, and he made the world finals team for the ACM International Collegiate Programming Contest.  No biggie. 

Tony's probably already blushing from me mentioning these things, so I will not continue on with the long list of accolades...except to say that he also runs marathons, plays the piano and violin, and does origami. I stopped believing he was a real human being a few hobbies ago. Oh, and did I mention he does break dancing?  We're eagerly awaiting this year's talent show to see Tony strut his stuff... 

In case you, too, don't believe he's real, here's a picture of Tony. I also pinched him to ensure I wasn't talking to a hologram.  

Side note: Tony notes that he runs half-marathons. No false claims from this guy.

So, what is your background? How did you find your way to Khan Academy? 

I grew up in the Chicago suburbs and went to the same high school as Stephanie - after that, I went to MIT and got the exact same degrees as Sal - and now I’m here! Must be destiny, I know. Anyway, I spent a couple years working in the secretive world of high-frequency trading. That was fun, but I’ve always wanted to understand how people learn. I taught classes for math olympiads and I’ve been fascinated by the “aha!” moment when you understand a new idea.

  Speaking of learning, that’s mostly what I’ve been doing at Khan Academy on the data science team. Well, learning about learning - I’ve been working on metrics to help us understand if changes we make to the site translate to users learning more over time. There are a lot of forces at play, many of which are actually psychological in nature. It’s been an interesting journey so far.

  You seem pretty good at programming...how did you get started?

  I was really into math competitions back in high school and I signed up for TopCoder my senior year since a lot of the same people were participants. I jumped into my first competition not really knowing how to program and not knowing what a “class” was. Yet, all the problems asked you to “write a class that does X” in order to solve the problem. Needless to say, I didn’t even come close to getting any of the questions right. That probably wasn’t the best way to start in retrospect, but it really motivated me to get better.

  The nice part about most programming contests is that you can learn from everyone else’s code afterward! The not-so-nice part is that I saw some pretty obfuscated C++ code as a first-timer, with lots of macros, gcc-specific extensions such as “a <?= b”, etc, you name it...

  How has your KA experience been so far?  

  It’s been great. Khan Academy is by far the most open environment I’ve worked in. At a personal level, I’ve always optimized for a combination of “how interesting is the work?” and “how much am I learning?”. Khan Academy is at the perfect sweet spot.

    Now that I've embarrassed you by outing all your hobbies and accolades, what's something random you'd actually like to share about yourself? 

  Oh, I designed a shirt for the Chicago-area ARML team back in 2006. We had 50+ people wearing it! I put this on my college application too, so maybe this is what got me into MIT. This would have been perfect for my Khan Academy application, but I totally forgot about it when I was applying!

Seems pretty appropriate. 

[winstonbinch]

A producer's definition of crowdsourcing

I’ve spent most of my career working as a producer, breaking ideas down to their core, figuring out what they are, and how they should work.

Over the past year, I’ve read and heard a lot of talk about crowdsourcing, and it’s mostly questions about what it means for the soul and future of our industry. What I haven’t heard a lot about is what it actually means. Blame it on a good buzz word, but everyone seems to be caught up in the effect rather than the what. As a producer it’s hard for me to skip the what.

When you break crowdsourcing down to its essence, it takes on a relatively familiar shape. It’s Web 2.0 + freelance, or more specifically, aggregated, online freelance. The only thing that’s new about this idea are the websites and tools that filter and bring the communities together, though a site like TopCoder has been around since 2001.

Crowdsourcing sites make it easier for small businesses, brands, and agencies, to access and activate freelance communities, at lower cost, and that’s a very cool and powerful thing, but even great tools don’t remove the baggage that comes with having to lead and manage a freelance and virtual workforce.

The best things we’ve done since I’ve been at CP+B, particularly in the digital space, have been a result of co-located and cross-disciplinary teams working together. But even the successes have come with challenges, and it can be hard getting people to work together well even when they sit next to each other in the same building. When you add physical distance, different time zones, tight timelines, and a lack of shared work culture and goals, it gets a whole lot more difficult to do effective, press worthy work. It’s not impossible, but it takes more time, and in the marketing world, that’s not something we have a lot of.

Crowdsourcing has a future. There’s no doubt about that. It’s a smart way to open up the ideation and design process, as well as empower consumers to take a more active role in your marketing efforts. But as good as the Web collaboration tools are today, the toughest creative problems, more often than not, get solved faster and more efficiently by teams sitting and working together.

For that reason, we’ll continue to experiment with and incorporate the latest and greatest collaboration and social media tools into our work, but as far as our business model goes, the premium, for the foreseeable future, will be placed on creating an evolving work environment and process that enables better teamwork. Original A-team style.

That’s how we get results. And that’s what our clients are paying for.

[cabbagebot]

Can't Decide

Whether or not I should wake up for the Top Coder competition tomorrow. On one hand, it shouldn't be difficult for anyone to wake up at 11:30 AM.

On the other hand, goddammit is it hard.