
Data Idols and our partners at the Data Science Festival are fortunate to know some incredible leaders in Tech across the world, and have even heard a few of them present at our events.
Tom Matcham, Head of Data at Goss, has given us some insights into his life, career & everything tech! 🚀
I have been working in machine learning and data science for around 11 years. I started my first company, a machine learning startup that helped games developers designed personalised user experiences, whilst at university. After a few years of working with games companies that company metamorphosed into another startup called TAZ Analytics that aimed to automate product analysis methods like funnel analysis and churn forecasting. Whilst neither of those startups were massive commercial successes, I honed my skills and discovered what works and what doesn’t. I now am Head of Data at Goss Media, a startup that develops a prediction platform for social media, TV shows and more.
I fell into it! Whilst at university my plan was to continue in academia. I had an offer to do a PhD in mathematical aspects of quantum mechanics but after I won funding for the idea that became my first startup it felt right to put the PhD on hold. I have no regrets about my decision – I’ve been able to satisfy my mathematical curiosity along the way and learnt many more skills than I would have done otherwise.
Demonstrable, quantifiable outcomes – ‘Ye shall know them by their fruits’. Data science as a separate discipline from statistics has developed in a very fortunate economic environment, one that has allowed companies to invest in data science hoping that it would be the goose that laid the golden eggs. If the good times are over, the teams that can’t causally relate their efforts to positive outcomes will have to rely on other skills to survive.
Best: finding counterintuitive results
Worst: pushing back against the ‘pop-science’ presentation of data science promoted by marketers – ‘data is the new oil’, ‘you can find insights if you look hard enough’ etc.
Make sure your statistical knowledge is very strong – if it’s not, you will likely make expensive mistakes. Read up on extreme value theory, fat tails and causal inference. Think very hard about the distribution of your data. Don’t do data mining.
By acting like a kid with my son. I take playing with trains very seriously.
Helping Goss to become a mega success. I feel like I’ve achieved more with Goss in a few months than I have in my entire career. That might sound bad but I needed those experiences to become useful.
I’m very happy at Goss but one day I’d like to build the world’s fastest data engineering pipeline.
I love programming and think our current data engineering solutions are massively over-complicated, leading to frustration and wasted money. Solving that once and for all would make me happy for the rest of my career.
– Python will continue to be awful.
– Data science lay-offs will follow the layoffs in software engineering.
– The hunt for AGI will continue to be a waste of time.
– Charlatanism may stop being the most attractive way to forge a career in data science.
– Prediction will continue to be near impossible.
Techies and non-techies welcome
for signing up to the Data Hub 🙏
You’ll get the latest news about our events, meet-ups, tools and resources straight to your inbox!