2016: A Life Update October 2016

2016: A Life Update

I keep telling my friends that I will give them an update on what I’ve been up to for the past few months. So here ya go. This past January and February I did a two month trip around Malaysia, Thailand, Indonesia, Singapore, Athens, and San Fransisco. While on this trip, I made the decision to study data science and make a full career transition. I started the Data Science specialization on Coursera from Johns Hopkins in February and quickly got to studying. That was going well… I was quickly becoming competent in R, and had a plethora of data science freelance offers from startups (more on that later).

By day I was studying data science, by night I was taking MMA classes at Williamsburg MMA and immersing myself in a new group of inspiring people through a group called Salon (which also tend to intermingle with a few other awesome circles). This led me to discover even more inspiration through the Met Media Lab. Quite a few friends were made through these groups as well as an overwhelm of feedback on my career and personal exploration.

But then in April, life happened and through a series of fortunate events, my glorious side project, DoubleBlinded, got a burst of media attention and my cofounders and I were (A) being interviewed a few journalists, (B) invited to talk on the Useful Science podcast (C ) getting flown out to San Diego to give a talk at an open innovation summit by MIT and (D) getting interviewed by Y Combinator. I spent three months working full time on the company — speaking with investors, studying the FDA, and seeking out advisors. But after three months of research, our experiment came to an a soft pause when our advisors and legal team advised us not to continue with our current model due to legal and ethical reasons as well as concerns about size of market.

Upon putting DoubleBlinded on pause, I decided to go back to freelancing to earn a living while sorting out what I really wanted to work on. I experimented with ideas of starting an outdoor trip leading company, doing biofeedback, starting an online school, becoming a quantified-self-based lifestyle/fitness coach — I’ve always wished I had a better way to utilize my knowledge of nutrition and physical chemistry, doing virtual reality (eg 3D modeling and unit development), and starting a software consultancy. And to this day I can say I am still very much interested in all of the above.

But what is interesting and what is executed on are very different things. I found myself a freelance gig in NYC — the highest paying gig I’ve ever had to date — and learned React + Redux over the summer while subbing for an Intro to Web Development class at the New York Code + Design Academy. This was my first time teaching and it was quite enjoyable, so this ended up being a successful career exploration. I hope to teach again soon.

Finally come August I decided to make time for some trips that I had been planning. I had a list of cities I had been wanting to see for ages, so I decided to cross a couple of them off — Portland and Montreal — as well as visit some friends in Boulder who I hadn’t seen in ages. Portland was a reconnaissance trip to feel out if it would be a nice place to live someday when I’m grumpy and tired of living in New York City. Montreal was a vacation. Around this time I learned that Hacker Paradise was going to Jeju Island in South Korea for a month, which fit my timeline perfectly, so I took the opportunity. I had met one of the founders, Alexey, about a year prior and been sold on the concept of their program. At the same time, I found a roommate and managed to find someone to sublease my apartment for almost the exact dates I would be gone, so there was no way I couldn’t take the opportunity.

And this brings me to where I am today: I’ve been in Jeju for two weeks. As my trip to Jeju got closer, I started to feel a hesitancy around coming here because I wanted to be productive and work towards my career goals instead of goof off and travel, but with the trip already booked, I did it anyway. Upon getting here I was pleasantly surprised with the quality of the coworking space (Jeju Space), the intelligence and motivation of the group of people I would be traveling with, and the overall setup of the situation for focusing and getting work done. I’ve been more productive while here in Jeju than I ever have been at home in New York City.

Since arriving to Jeju I parted ways with my freelance client and have taken to studying data science full time — I’m spending about eight hours a day learning machine learning, statistics, and data cleaning, this time with a python and javascript stack (Pandas, Numpy, Scikit-Learn, Rodeo) and will soon be learning Spark.

I’ve also taught myself a few new things about Javascript <warning: I’m about to geek out>… using ASM.js to write C code that compiles to highly optimized Javascript and produces code that runs at around 2x native performance. Also SIMD to do mathematics in vectors of 4 to get 4x performance over traditional computations. Holy shit! There’s a saying in the Javascript world that right when you think you know everything about Javascript, you are the most wrong. (Is that a saying? If not, it should be. I believe I might have stolen this from my friend Richard Jones.) Eventually I hope to see data science tools like those available for Python existing for Javascript. I’ve actually been prototyping a Numpy-like API in Javascript while learning actual Numpy… similar to Numjs but with a syntax closer to Numpy.

While studying data science both this spring and this fall I’ve been periodically reaching out to interesting companies to see if they’d be willing to employ me while I study under the understanding that I’m still learning. The general response I get is pure excitement to find someone that is both a good programmer and a data scientist. The companies I’ve spoken with so far generally are able to find good engineers and good scientists, but not both. By the time I’m done with my studies, I’m excited to be able to say I am both.

So what is my motivation with data science? I’ve become increasingly bored with web development and engineering over the past years. I began to feel that I wasn’t solving new or exciting problems and was instead simply getting paid to implement some of the same old patterns over and over and over again. Over the summer, I hoped that learning React and Redux would re-ignite my excitement for front end development, but this was simply not the case. I found ES6 (the latest version of Javascript) to be phenomenal and fun to program in, and I find React and Redux to be major improvements in the way that modern front end development is done, but I simply can’t say that I found excitement in them. My motivation with programming has always been purely for the product that I was building and never about the process of building things themselves. Turns out this is a bit problematic when you are getting paid to implement code for companies whose domains you aren’t really excited about.

Additionally, I’ve been increasingly aware that my job as a software developer is getting easier and easier to outsource to talented developers outside of the United States that are able to work for much lower wages than I am interested in working for because of lower costs of living. Sure, this isn’t the case just yet, and I don’t think my job will ever be completely replaced by outsourced workers, but I do think that a lot of the things I was doing before will be. So instead, I desire to solve mathematical and scientific problems artistically. I’ve always considered myself an artist far more than anything else, and this is these are the roots that I hope to get back to. I hope to whatever god might exist out there that one day I can be a data artist moreso than a data scientist. I’m inspired by my predecessors like Mr. Feltron and exhibits like teamLab.

My first and most exciting data science project so far has been doing emotion analysis of myself and my friends over time. My friend and first research subject, Stan, built an app called Lifeslice about five years back which is a simple menubar app that runs in the background and takes pictures of you every few minutes while your laptop is on. He and a small handful of other people on the internet have been using his app for the past few years and have massive datasets of photos — he recently sent me 15,000 photos taken over five years — enough to pick up actual signs of aging and potentially significant life events. I’ve been taking these photos and running them through a few emotion detection algorithms including one by Affectiva and one by Microsoft. I’ve then been taking these emotion scores and combining them with sentiment analysis from iMessage, journal entires, and Facebook Chat as well as movement, weather, and location data to find correlations. The goal is to build a predictive model to show how these variables relate and create the ability to show changes in emotion over time. — for example, the ability to show the effect of a significant life event such as a death of a loved one, a marriage, moving, or starting a new job on someone’s emotional and physical health and quantify it.

So the takeaway of 2016 thus far as been that I’ve crammed many years worth of travel and learning into nine months of time. I’ve not found much desire to write in the journey as I am continuously humbled by the existing piles of writing on everything that I’ve taken to studying that already exist. One of the interesting things about personal exploration is that in trying new things, you are repeatedly a beginner. I desire to become known for something and to be the best in the world at something, and it is when I do this that I will find the most motivation to write. For now, sure, I have things to say to my friends, but I have little wisdom to contribute to the greater community. Yet. That will change.