How Sandtable's cloud technology changes data sciencePosted almost 3 years ago by Michael Oliver

Dr Thomas French is the CTO of Sandtable , a data science company in London. In this edition of The Story, He tells Client Server’s Michael Oliver about they use their tech for modelling human behaviour and more.

Michael Oliver: Tell me about Sandtable.

Thomas French: We're a data science company. We try to help people solve their problems through analysing data and building simulations, which are virtual laboratories where they can run experiments without having to run them in the real world.

MO: What kind of tech do you use to achieve that?

TF: We’ve developed a cloud-based platform to facilitate our work. When I joined the company eight years ago, that was part of the challenge. The cloud was is this emerging technology that we figured could help us run large-scale simulations more effectively. We develop with Python and deploy using Docker on Kubernetes; it's a very modern, agile stack.

MO: Can you describe the kind of organisations Sandtable works with?

TF: We’ve worked with many different types of organisations across many sectors over the years, but now we specialise in two main areas: marketing and public policy. Typically, we work with large enterprises and some government departments and universities.

MO: It strikes me that a stack like Sandtable’s would require a certain type of individual?

TF: We're quite an academic group; about half of the team have PhDs, mainly in the data science team. For engineers, we look for developers with a strong CS background.

MO: How would you describe the culture at Sandtable?

TF: We are a company dominated by introverts. We're quite a deliberative and academic crowd, but also collaborative. It's been an ongoing process figuring out how best to collaborate given different types of personalities.

MO: Are you the type of CTO who still likes to code himself?

TF: Very much so. We’re a small company and I still get my hands dirty. I wear many hats.

MO: How would you describe your team’s approach to work?

TF: We are lightweight in our development processes. We’re not wedded to any particular agile methodologies; if we see a need for change, we switch things up. The main thing is to remove barriers – let people get on with their work.

MO: How would you describe your relationship with Client Server?

TF: We have a great relationship. They’ve done a great job for us in filtering and managing the hiring process, and we’ve had a lot of success working together. Client Server understands what we’re looking for and understands us as a company – that’s massive.

MO: What are some of the big trends you see coming in the data science space?

TF: The really cutting-edge stuff is in reinforcement learning. We’re looking for real use cases where we can apply it to our models.