I Left Financial Tech… Here’s what I do Now

This post is long overdue. I actually left financial tech 3 months ago, but I’ve been caught up in a number of projects. The more interesting your life gets, the less time I have to blog. I’m trying to find the time by cutting down on Netflix time. Reading is now done entirely on the tube, and walks are accompanied with an audiobook. Coding is now every day including the weekends. Every now and then I will drink alcohol with friends at the end of the week, but this is always met with regret the following morning as I’m now 30, and it takes a day for the mind fog to fully clear inhibiting my ability to write code. I’ve also been writing technical guides which can be found on my medium [link]. What I’ve described sounds like hell to some, but it gives me joy. I have been given opportunities that I never thought I’d have, met interesting people, been at some pretty cool places, and worked on some projects that utilise future technology. I’ve been able to peer under the hood of systems that keep the world ticking over, and had the pleasure of coding code reviews of other developers. With code reviews, it’s like getting a snapshot of how someone else thinks. The love of code, the powers it’s given me, the ability to solve a range of problems at scale, and the life style its provided means I don’t have to think twice about if it’s worth it.

So what’s my day job now? You guessed it, more software engineering. This time however, it’s in the mechanical engineering fields. My open source deploy-ml package got the attention of some guys at the startup Monolith AI [link]. I now work for them. I like AI but anyone who knows me knows I’m very wary of the hype around AI. My deploy-ml project wasn’t about pushing the most cutting edge AI, but making the training and deployment process of standard machine learning algorithms more stable. This is what got me excited about the startup. I get contacted multiple times by people with super ambitious AI ideas, and the news is awash with companies claiming to replace humans, predict the future, or do things that have never been done before. Sure I guess some of them might achieve that but I don’t really like their chances. Monolith’s aim offered real tangible value. It uses AI to speed up the design process.

For instance, let’s consider a wind tunnel test for a car. There’s different configurations for a spoiler, so what engineers do is test all those configurations in wind tunnels. This is costly and time consuming. What monolith does is test a sample of configurations, then get the machine learning to fill in the blanks:

Screenshot 2019-09-14 at 15.08.02.png The engineers then have a detailed map of all the outcomes with just a few tests. They still test their final design decision, it’s just costs less money and is quicker because not every wind tunnel test is needed for the engineer to make their final decision and test it. This practical AI approach with clear benefits was easy to get behind. It’s already, made gains in car, aerospace, and packaging where it saved 2 months in product development and reduced the amount of plastic in the packaging by 40%. We’re now ramping up, we made Forbes as one of the UK’s top 29 most exciting AI startups [link].

Can’t wait to see how this new adventure unfolds. It’s got me running on all cylinders but it feels good. I personally feel that the flexible office hours enable you to ride grooves when you’re on a roll, and recharge when you’re not.

 

personal mutterings

maxwellflitton View All →

I help clinicians get to grips with coding and tech, I also code for a financial tech firm

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: