On January 12th, 2012 I went to Airbnb to give a talk about Storm at their tech talk series. It was a public event and quite a few people came. As usual at these events, there was mingling/socializing after the talk. A bunch of people approached me and one of these people was Jake Klamka.
There's nothing striking about Jake Klamka when you first meet him. He's a pretty unassuming fellow. So when he approached me after my talk, little did I know I was about to meet an astonishingly good entrepreneur – someone who's a master at getting help from others. Over the next few years, little by little, he would get me to help him more and more. And I would be amazed at what an awesome company he would build.
People approach me all the time with various requests. And the vast majority of them go about it the wrong way. For example, I often get emails – from complete strangers that I don't know – asking to meet with me for an hour or have a conference call with me. The reason they want to meet is typically vague – "I want to find out more about Storm" or "I'd like to get your advice on dealing with Big Data". In the past I have taken some of these meetings, and they are almost always a waste of time for both of us. They typically come to me with generic questions that could easily have been answered by doing a Google search, reading my blog, or watching one of my talks. I want to help people, but coming to me with questions that you could easily have answered yourself with a slight amount of research is not productive.
Jake puts a lot more thought into how he asks for help. A lot of what I learned about getting help from others I learned from how Jake slowly reeled me in to helping him. When he approached me at the Airbnb meetup, he didn't start with what he was working on and what he wanted from me. On the contrary, he asked about me and what I was up to. He told me he'd been following my work for awhile and asked me how I liked working at Twitter post-acquisition (I was part of BackType which was acquired by Twitter 6 months before). He asked me about how I came up with Storm. He did not attempt to shift the conversation to himself at all. A clever fellow, this Jake.
Naturally, I felt compelled to know more about this person who was flattering me by showing so much interest in my life. So I asked him what he was up to and what he was working on. He smoothly described to me how he had gone through Y Combinator but was now working on something new. He told me about his idea to form a fellowship program (which he later called Insight Data Science) to train PhD's with physics, math, and other backgrounds the skills necessary to become data scientists. The idea was that the PhD's were a surplus of very smart people without the greatest job prospects, and at the same time there was huge unsatisfied demand for data scientists in the industry. The PhD's already had the statistical skills necessary for data science – they just needed to learn the tooling of the industry. It was a smooth pitch and I told him I thought it was a pretty interesting idea.
In retrospect, I believe that he came prepared to pitch me on this idea, to gauge my interest, and to see if he could get me involved in helping out with the program. But here's the key: he didn't force himself on me. Rather than approach me and pitch me, he took advantage of natural social dynamics and waited until I asked about him. Not until then did he give his pitch. And not until I told him I liked his idea did he make his ask.
Jake asked if he could come by Twitter for lunch one day and go over his plans for Insight. It was a minimal commitment and one which he made as convenient as possible for me (by coming to where I work). Unlike the requests for 1 hour meetings I get from completely random people, this was appealing because it was a request from someone who had demonstrated he was doing something interesting and who was asking for specific feedback. So I was glad to take the meeting.
Jake came by Twitter for lunch and was super prepared. He brought a slide deck on his iPad and very clearly took me through the specifics of how Insight would work, how he was going to recruit students, and how he would connect them to industry. By making everything so tangible he made it incredibly easy for me to give him feedback. I told him it was important that he get companies involved in the program from the start because of how crucial it was to keep the students connected with the needs of industry. It was an extremely productive meeting and I felt like I provided a lot of value. I was impressed with Jake and outright told him I'd be happy to help him more in the future.
Over the next few months I helped Jake out with a few things – mostly related to connecting him to people within Twitter who might want to hire out of his program. That summer he launched the first batch of the program and asked if I would come down to Palo Alto to give a talk to the students. I was happy to do so.
That went well and the first batch of the program went on to be a huge success. Insight achieved – and continues to achieve – a 100% placement rate for all its fellows. Fellows get hired at all the top companies like Netflix, Twitter, Facebook, LinkedIn, Square, Microsoft, etc. It's a really impressive program. The most mind boggling part about it is that it's completely free for the fellows.
Jake and I occasionally kept in touch over the next year and a half. Then this February, he emailed me to update me on the program (it had grown a lot) and ask if he could meet with me again to talk about his idea for a new data engineering program. Unlike data science, this program would be geared towards people who want to build the pipelines underlying data products and the infrastructure that data scientists use. Like usual, Jake made the meeting as convenient as possible for me by coming to meet me near where I live in San Francisco.
I was skeptical about the program at first. While the data science program took advantage of a distinct imbalance – a surplus of smart PhD's and a deficit of data scientists – I wasn't entirely sure who were the target candidates for data engineering. However, as we talked I realized what a good idea this was. Data engineers are in huge demand, and getting skilled at data engineering is a ticket to a job that's not only high paying, but also supremely interesting – imagine playing with the datasets of Twitter, Netflix, Spotify, or Khan Academy. Everyone seems to have a "Big Data" problem now, and data engineers are crucial to solving those problems.
Besides that, it also became apparent that Insight is valuable in and of itself, not just as a ticket to a job in industry. Jake has created a cradle of creativity – the entire program is structured around fellows doing self-directed projects and helping each other learn and grow. The data engineering program seemed like a fantastic opportunity for any programmer to sharpen their skills and get an awesome job, whether a new grad or someone experienced looking to transition their career.
In our meeting Jake's biggest question mark was what kinds of people would make good fellows for the data engineering program. I felt strongly that you won't turn a non-programmer or an inexperienced programmer into a good programmer in 6 weeks, so the candidates he should go after should already be good programmers. The program should be focused on teaching fellows the tools and techniques of data engineering, not on teaching them basic engineering skills. This was a big difference from the data science program, and Jake and I spent a long time talking about how the interview process for Insight Data Engineering should work.
A few weeks later, Jake asked if I would become an official advisor to Insight. I accepted instantly. I found Insight to be hugely impressive, and Jake had long ago proven to me how resourceful and effective he is. Most importantly, it's always a joy to work with Jake because he makes sure that every meeting is productive. He asks for specific feedback and never asks overly vague questions. As an advisor to the program, I'm helping fellows with their projects and holding sessions to teach fellows various aspects of data engineering (like Lambda Architecture and Storm). I'm excited to see the program improve and evolve as it takes on more batches of students in the future.
It's fun to look back at how Jake reeled me in to helping out with Insight. He didn't just come up to me with a list of things he wanted – instead he first built my interest in him and what he was doing. Then he gradually escalated the commitments he asked for – first a lunch meeting where I work, then some introductions, then a visit to the Insight office, then help with evaluating candidates, then a regular commitment to spend time with the fellows. At each stage he made sure I was sufficiently interested such that taking on the commitment would be something I'm excited about – and not just a favor. That's what makes Jake such a great entrepreneur – he makes people who were total strangers want to go out of their way to help him. I think everyone can learn from Jake's example, as I certainly have.