Meet Pat, IRIS' Lead Android Developer, the man behind the wizardry of our model naming conventions and a part of some of the magic that happens here at IRIS.
What is your role at IRIS?
I've been at IRIS for 2 years now — I joined in August 2020 right in the middle of lockdown. The job ad I answered might as well have just said "Pat wanted" — someone with a music tech background and plenty of recent experience in Android Bluetooth development. Initially, I was working on the IRIS Listen Well app for Android.
My main focus at the moment is working on our cross platform SDK. This is a single codebase that runs all our magic on Windows, MacOS, and Linux, and powers our desktop IRIS Clarity apps. I also work on custom integrations with our partners. My ultimate goal is to enable IRIS to run on as many platforms and in as many ways as we need. The list is constantly growing which keeps me busy and out of mischief. Most of the time.
A typical day involves coding in either C++ or Python, a bit of DevOps to support all the different build configurations we need to maintain, benchmarking to highlight areas we can optimise our code, and listening tests / audio analysis to check we have optimal audio output.
You’ve got to create several products now for IRIS. What’s your favourite part about this process, and what is the most interesting thing you’ve learned along the way?
It's always immensely satisfying to see and be part of an initial spark of an idea growing and transforming into a released product. My absolute favourite part of all of this though is what I call "new model day", when the research team sends over a new model for me to integrate and I get to run some audio through it and see how it sounds. I also enjoy the close collaboration between the research, product, and marketing teams to bring the whole thing together.
Something people may not know is that we name all the different models we develop for IRIS Clarity. You’ve been the main driver behind our naming convention. Can you explain what the model actually is, and share what inspired its naming?
When we set out to build IRIS Clarity we were doing a lot of experimentation, trying to find the best balance between model architecture, audio quality, and performance to end up with the real time solutions we use today. New models were coming in thick and fast from our incredible research team, but they had names like "model-20210813-step001320000-66" or other such gibberish. It quickly became apparent that it was impossible for the team to have a rational conversation about the differences between two models without everybody asking which model we were talking about every 5 minutes, so we needed to come up with codenames for them all.
I realised the answer was right in front of me on my shelves. As Tony Wilson once said, "Manchester kids have the best record collections" — so all of our models are now named after bands or musicians with (sometimes vague) links to Manchester, where I am based. It's a lot easier to talk about the differences between bez or curtis or elkie or harper for example, and there is a never ending list of codenames to choose from.
As to what the models actually are / do — in a nutshell, they take a massive bunch of numbers and turn them into a massive bunch of different numbers. I'll leave a more in-depth technical explanation to someone in the research team...
Some quickfire questions:
App you couldn’t live without: Definitely Bandcamp. Hi-res audio in a single click, artists get more money, and I get a record in the post a month later unexpectedly when I've forgotten all about it. It's perfect.
Favourite noise: My next door neighbour (now sadly passed) used to spend all day in his garden sitting in his shed listening to cricket on the radio at a volume just too quiet for me to hear properly. I'm not a cricket fan at all, but the low murmur of the commentary, an occasional thwack of a ball, and a smattering of applause on a sunny summer afternoon is incredibly relaxing. So, the sound of someone else listening to the cricket is probably one of my favourite noises.
Biggest pet peeve: Litter. Pick up your rubbish and take it home if you can't find a bin. It's not hard.
Best concert you’ve been to: Oh that's a tricky one. Too many to pick from. James at Alton Towers? Flaming Lips at Jodrell Bank using the Lovell Space Telescope as a giant projection screen? Richie Havens at the Lowry? Hallucinogen in the middle of some field in Luton? Honestly, I couldn't pick one.
Something your colleagues don’t know about you: I used to live an incredible double life — green grocer / dry cleaner by day, sound engineer in a comedy club by night. I could probably tell you what all the cleaning symbols on your clothes mean. "This shirt is dry clean only. Which means it's dirty” — Mitch Hedberg.