IFTTT was the first of a few VC-backed startups I joined on purpose. I wanted to learn how they operated from the inside, and this was a good one to start with: a product everyone had heard of, right as the "that" in "if this then that" started to mean physical things.

I built 15 hardware channels in six months, each one a partnership: spec the triggers and actions with the device maker, build against their API (sometimes while they were still writing it), and coordinate the launch. Nest, Parrot, Withings, Fitbit, Belkin, Netatmo, Automatic, Lockitron. Platform IoT usage went up noticeably. On the side I replaced the aging Sphinx search with Solr on AWS.

It was also the job where my hobby projects and my day job were the same thing, which I recommend.

Background

IFTTT's whole pitch is "if this, then that," and in 2013 the interesting "this" and "that" were starting to be physical things: thermostats, fitness trackers, light switches, door locks. I joined to build that side of the platform. Head of Hardware sounded fancy; mostly it meant a lot of API docs and a desk covered in gadgets.

What I did

I built 15 hardware channels in six months, which took platform IoT usage up noticeably. Each one meant working with the partner to spec out the triggers and actions, building the channel against their API (some of which were being written at the same time), and coordinating launch with our team and theirs. Partners included Nest, Parrot, Withings, Fitbit, Belkin, Netatmo, Automatic, and Lockitron.

Away from hardware, I replaced our aging Sphinx search with Solr, deployed with Chef on AWS, and did the usual improvements and bug fixes on a large Rails front end with a very large user base.

Things I learned

Every device maker's API has a different idea of "state changed." The channel abstraction has to absorb that.
Hardware partners ship on hardware timelines. Software is always waiting on a firmware build.
This was the job where my hobby IoT projects and my day job were the same thing. It's a good place to be.