nawl

AT&T flew me to New Jersey to interview for their grad school fellowship program.

On that note, I've noticed that the one comment everyone makes when I mention how much I've been flying is "Racking up the frequent flyer miles, eh?" It's probably only the sort of thing I find irritating when I'm jet-lagged.

Kevin left early Tuesday morning. He was either sick or jet-lagged for the entire visit, so I spent all my time caring for him. And I came down with the cold the day of my flight.

I left Thursday morning and arrived in Newark Thursday evening, just in time for a lovely snow storm. The limo (actually a luxury taxi) was almost half an hour late from the storm traffic. The driver was hilarious, driving like a maniac in the snow, swearing at the other drivers, and at one point he even drove over the snow-covered curb dividing the express portion of the highway from the exit portion to get past some drivers who were going too slow for his taste. Fortunately we actually made it to the hotel. And man, what a hotel. My room had a full kitchen. The bed was wider than my bed here is long. I took a bath.

The next morning, I woke up at 4, then 7:30. At 8:30, the fellowship's academic advisor called me for a phone interview. At 9, I was driven to the AT&T research buildings to begin the real interviews. Everything was covered in fresh snow. The AT&T buildings seem to be in the midst of a semi-abandoned corporate park, the victims of market changes and downsizing and companies building buildings they don't actually use. There's a statue of Shannon in the balcony, and pictures of various other research giants other places in the building maze. I had no idea where I was the entire time I was inside. It was nice, though. The cafeteria has wide windows looking out onto a peaceful snow field, and at lunch we watched deer with big white tails bound back and forth.

The interviews themselves were quite fun, actually, a series of 45-minute sessions where the people "interviewing" me actually spent most of their time telling me about their big projects and then asked me for questions about the program. I 1. learned about effectively sampling IP routing data, 2. got grad school advice and criticized the W3C, 3. got my first dose of approximating the TSP, 4. discussed compounding engagement cheesecakes, 5. learned about optimizing compression algorithms for phone call data, 6. learned that math is a really, really small world, that the chromatic number of a distance-one graph over R^2 is (less than or equal to) 7, and saw how to look backwards at the prize-collecting Steiner tree problem to approximate it, 7. learned about computing exact values for large problems of the TSP and saw more Rubik's cube variants than I've ever seen in my life, and 8. got another dose of TSP and had an unexpected meeting scheduled for the next day with a Princeton professor.

But really, the most awesome interview was with Neil Sloane. When I walked into the office, he pointed to a bookshelf a few feet wide and told me "Those are my books." as in, the ones he wrote or co-authored. I told him I was a fan of the OEIS, and began to describe how I used it to solve Kevin's Christmas present puzzle. I got as far as describing a design of paper folded at right angles before he said "Oh! The dragon curve." So then he told me he spends most of his time updating the database by hand, and showed me some of the most recent update suggestions in his email. Then he showed me a series of problems inspired by silly integer sequences, investigating an operation invented by Knuth involving writing a number as a sum of terms of the Fibonacci sequence, and looking for primes and squares in "degenerate math", where addition is max and multiplication is min and you carry out multi-digit multiplications the way you learn to in third grade. Then he went on to a problem of cutting up a polygon into pieces that fit into a square and its equivalent in three dimensions (which I mentioned reminded me of the Erik Demaine video of cutting and folding the surface of the cube), and had just started talking about error-correcting codes over algebraic rings when we ran out of time. This was all presented at light speed, in the nicest most soft-spoken friendly manner possible, and I was struggling to keep up the whole time. Fabulous.

I also had an hour-long interview "tea" with the members of the fellowship committee from other (non-math) areas of AT&T. It was tele-conferenced with another building somewhere else in New Jersey, so not only was it an eight-on-one interview with jet-lagged me near the end of the day, but I was also on camera. By the end when they asked me if I had any questions for the committee, I'd managed to exhaust all my questions on earlier interviews, so I asked the representative from natural language processing what AT&T's interest with natural language was. The answer was "We're a communications company, of course we're interested in natural language!" But then it devolved into a rather amusing discussion of the AI phone service being put into place for customer service, and of the records of people who call it who either don't realize they're talking to a machine or who immediately realize they're talking to a machine and are incredibly unhappy about it.

I also heard quite a lot about budget problems, disappearing funding for summer interns, the shrinking of the fellowship program, journal subscriptions that were first made electronic and then canceled, the unionization away of secretaries, and some amount of trepidation for the upcoming purchase by SBC and the future of research in the company. I was surprised to learn that there were actually only around 40 applicants for the program, of which 12 were interviewed for 5 slots. I don't think that's private information.

The official day ended when dinner ended at 8. After 12 hours j'etais crevee.

And so I spent the rest of the weekend with Christophe.

That isn't technically true.

Saturday afternoon, I got a personal tour of the Princeton campus, the IAS buildings, and the CS department from Moses Charikar, the poor CS professor roped into showing me around by his friend who interviewed me the day before. The Princeton campus really does look like a gothic church. I'm not sure how I feel about the possibility of living in a building that should be on Mont Saint Michel. I was once again shocked to discover that my application had been read in detail by an actual human who paid attention to what I wrote, recognized me as an individual, and who appears to believe I am interesting in some way. Later I talked to a graduate student who told me she was pleased to see that there would possibly be another girl joining the theory group. I think I'm beginning to be sold on it all. Small program. Tranquil atmosphere. Young theory professors working in young, sexy areas. Emphasis on close interaction with faculty. Possibility of getting an advisor from anywhere I want, including IAS or math.

Christophe took me rock climbing, where I took the belay class and was having fun until I strained my forearms on a tricky climb. We went out for sushi, where he was able to show off his new ability to eat with chopsticks. My French is *awful*. The next day, we drove to New York to see the last day of the exposition of The Gates in Central Park. Two hours there, two hours back, and a parking ticket because we didn't have enough quarters to fill the meter entirely. It was bitterly cold. The crowds were pretty impressive, and nice people with tennis balls on sticks drove around in golf carts and passed out small orange patches of fabric. The orange color was nice against the snow. I thought it looked like ski slalom poles, a view independently shared by quite a lot of viewers despite the fact that slalom poles don't have curtains on them.

I left late on Sunday evening. The weather was clear, and the takeoff gave me spectacular views of Manhattan. I even had a free seat next to me, so I slept fitfully curled up in a fetal position between the armrests. If there's one saving grace to such a short trip, it's that the whole time my body had no idea what time it was, so going east wasn't much of a change.