Years ago, I worked for a robotics company that got its start with funding from a small business grant from the Department of Defense. This was my first introduction to Small Business Innovation Research (SBIR) grants, which advance government research and development by providing funding to innovative small businesses. Each agency with a sufficiently large R&D program is required to set aside around 3% of its R&D budget for SBIR & STTR (tech transfer) grants, and NIH certainly falls into that category. These grants are split into two phases – Phase I for proof-of-concept and Phase II for commercialization – and require a proposal that is similar to what a PI at an academic institution might write. It’s a bit shorter, thankfully, but often the small business won’t have a grants office or dedicated personnel for the task, and that’s definitely the case here at FlySorter. I’d estimate that we spent more than 100 hours preparing our most recent proposal for submission (that includes registering for various government websites and completing forms, along with actually writing the document).
The application we sent to NIH in January of this year is actually our third try. The first time was a swing for the fences several years ago which received a mediocre score, the second was a more modest application in 2016 that came pretty close during review, and this year’s was a revised proposal (revisions are a special case that allow you to reply more or less directly to reviewers). We got our (very promising) scores in March, and had to wait another four months to get the official word: our application was funded!
Our grant comes from the Office of Research Infrastructure Programs (ORIP), and will greatly accelerate completion of our automated classification system. My brother Matt (subject of a previous post) is on sabbatical this coming academic year; the grant will fund a semester of his time and bring his considerable expertise in computer vision and machine learning to the project. It will also cover a full-time intern at our Seattle office, a portion of my salary, and a host of tools, parts and materials. And more than just monetary support, getting this grant also provides validation for the idea, something that can be hard to come by when you’re largely working by yourself and don’t have a ton of product on the market yet.
The first milestone we’re aiming for is to finish up the prototype image capture hardware (the Fly Photobooth), which will allow us to record tens of thousands of images of flies to train up our machine learning algorithms. Stay tuned!