What defines large-scale machine learning? This seemingly innocent question is often answered with petabytes of data and hundreds of GPUs. It turns out that large-scale machine learning does not have much to do with all of that. In 2013, Léon Bottou gave a class on the topic at Institut Poincaré. The class is still as relevant today as it was then. This post is a short summary of it.
The fundamental hypothesis of machine learning
Most of the recent progress in machine learning progress has been driven by the paradigm of learning by which we train a model ƒ from existing data. We estimate ƒ using…
So you’re becoming a tech lead. You have the support of your manager and your team. You will be getting formal training soon and your manager will be coaching you. This is all nice and great, but… what could help you become a great tech lead now?
This post aims at helping you get started and being the best possible tech lead from day one. One of the main novelties that you will encounter as a lead is about people management. …
Apparently, software is eating the world and AI is eating software. As more and more companies are turning to machine learning (ML) to disrupt their businesses, it occurred to me that a split was appearing between companies that are truly ML-centric and companies that aren’t.
What defines an ML-centric company? What is the difference between a company using ML and a truly ML-centric company? An ML-centric company manages to put ML at the core of its activities. The way it drives projects and processes is just different. Having had the chance to work at such a company (namely, Criteo) over…
Going single-track
RecSys changed its format this year and decided to go single-track. I’ve always enjoyed single-track conferences a lot so I am definitely biased, but the single-track has a lot to say about it. One of the unexpected advantages of it is that it helps bring people from various backgrounds around the same problem/topic. This is particularly useful in conferences like RecSys where the background of participants is very diverse (covering algorithms, math, sociology, machine learning, data science, UI/UX, etc.). Another way of looking at this is that single-track forces the discovery of new topics. …
Many years ago, I was part of a team that built one of the first autonomous cars in the world.
The team was created to enter the now famous DARPA Urban Challenge.
We worked like crazy for 18 months in a row. Night and days. Week-ends included. We didn’t care. We were so passionate about what we were building that we did not suffer from it. We loved it.
At the end of the event came the time to give prizes to the finalist teams. We were one of them.
All teams were happy, but exhausted. These past 18 months…
Olivier Koch
machine learning / computer vision engineer & tech lead
