Week 4 has been all about demos. Polishing our demos, picking companies that we want to demo at (and possibly interview at later on). Every morning, we practice our demos, 10 minutes per person, with the goal of keeping our demo to under 5 minutes to leave time for Q&A. I've found that the act of rehearsing our demos makes it much easier to pick out where I need improvement. For example, I tended to have trouble with explaining the validation portion smoothly, even though I knew what I was doing there. A tool that seems useful, especially for short demos, is to write out exactly what I want to say, and that definitely helped.
On the type of work that I'm interested in, here's some things I've become much clearer on.
Firstly, the factors I'm considering for a company. The ideal combination is: a company that deeply values the hard sciences (in my case life sciences), and is solving very tough technical problems that requires growth in and mastery of deep technical topics, on a team that encourages experimentation, personal growth, and open source contributions on company time. We'd have to be at the innovation boundary of very powerful techniques. This is important for me, because I believe that 5-10 years down the road, I would have mastery over very foundational and broadly applicable tools with the appropriate experience applying them to real-world problems, which I could leverage to solve more cool and interesting problems. It's also a good defence against being pigeon-holed into a particular domains or tasks - autonomy in problem selection and definition is very important to me, so most of my choices aim to maximize that over money.
Secondly, I've effectively ruled out companies that are dealing with non hard-science data, e.g. insurance claims, marketing & advertising, finance, and business data. Having applied computation to the life sciences over grad school, and being trained in the life sciences for over 10 years, I'm not ready to give up that background knowledge to work on other problems. I also believe that investing in the hard sciences means investing in the next wave of real-world innovation, and I'd like to ride that wave.
Thirdly, within the next 5 years, I see myself growing as a technical person, rather than a management person. People issues, particularly conflict resolution, make it difficult to focus on being a good craftsman, and I much more enjoy craftsmanship than management.
Now, on the companies that have come by...
Most are using open data science tools in their toolkit, and this mostly means Python and R, Spark and a few other big DB tools. Some are still using SAS (.................) and didn't show a trend towards open data science languages, and effectively ruled themselves out of contention. (Using legacy tools signals a lack of forward-thinking and a desire to favour the status quo over pushing boundaries.)
Some have given us words of wisdom. One guy basically said that healthcare has messed up (he used stronger language) incentives. Another said that to solve healthcare we need to first solve human behaviour. All very interesting points that are well-taken on my side. A non-healthcare company told us that if we're not paying for a service, then we're the product.
In our session, it was basically the pharma research arms that piqued my interest the most, aside from one hospital's internal startup team. The gap in interest between #4 and #5 (for me, at least) was really big, and the gap of interest from #5 to the rest was even larger.
Anyways, week 5 begins soon, and we pivot over into interview prep. Looking forward to learning lots, particularly doing deep dives on my weak spots!