QuantumBlack’s Data Science Internship: Room to fail and an environment to thrive in

This summer saw four research students join the team for QuantumBlack’s inaugural Data Science Internship Programme. Here at QuantumBlack, diverse backgrounds, experiences and personalities are not only welcomed but encouraged. We built our intern team with this in mind, choosing from hundreds of applications to bring together talented PhD students from different nationalities and universities, all with strong mathematics skills and a passion for data science.

From the start, our participants were given the freedom to shape the internship to fit their individual expertise and interest areas. Tackling a variety of problems and sectors, from deep learning to crunching numbers on live client engagements, from pharmaceuticals to banking, the only ask we had for our interns was that they provide an insight for QuantumBlack to work with.

Our interns were granted complete freedom to shape their research at QuantumBlack.They were required to provide an extensive assessment on whether their project could bear fruit in the future or needed to be taken in a different direction. Everyone was given room to fail -a key ingredient for success when working with data. They were dealing with untried and untested research, and even if their project was unsuccessful it would tell us something we didn’t already know.

Alongside weekly check-ins with assigned mentors, our interns had access to a range of experts from across our business. If they wanted to work through a challenge with our Head of Data Science or an Associate Partner, or take on a more client-facing role, there were opportunities for them to do so.

We’ve been delighted with the results of our first summer internship programme and have benefitted hugely from the fresh perspectives. Details on how to join our 2019 cohort can be found at the end of this post -in the meantime, here are reflections from our interns about their time with QuantumBlack.

Roxanne Zhang

A code ninja from China who loves pigs and memes!

What’s your background?

I am pursuing a PhD at Imperial College London, where I have worked on applying deep learning to predict traffic flow using GPS trajectory data while considering the data sparsity problems associated with GPS trajectories.

What was your focus during your internship?

I worked on applying Bayesian Networks (BN) to the banking industry. I had the opportunity to work closely with other data scientists, data engineers, analytics engagement managers and machine learning engineers. My primary responsibility was to add additional features to an internal R&D BN python package and provide fairness analysis based on counterfactuals.

What did you enjoy during your time at QuantumBlack?

The internship presented an opportunity to greatly improve my technical skills, and I seized my time at QuantumBlack to work with algorithms from recently published papers, coding and applying them from scratch. I was also able to develop my soft skills. Standout moments for me include presenting 12 weeks of hard work to a room of data scientists, listening in on problem-solving sessions and receiving feedback directly from QuantumBlack’s Global Head of Data Science.

Marc-Andre Schulz

A pleasant scientist who loves plots and is very hard to annoy!

What’s your background?

I am currently a PhD student at RWTH Aachen University in Germany. My research focuses on semi-supervised deep neural networks and their application in brain imaging. I exploit general purpose neuroimaging databases to improve prediction performance in small sample biomedical datasets.

What attracted you to QuantumBlack?

This internship seemed like a great opportunity to see data science applied at a much larger scale, as well as a chance to experience how industrial research differs from academia.

What did you work on during your internship?

During my internship I worked on ‘heterogeneous transfer learning’ -the attempt to transfer model-knowledge from the original feature space (known as the source domain) to a new and different feature space (or the target domain). I evaluated approaches to improve transfer learning in fully connected neural networks and tested adversarial training procedures to condition the target domain latent-space on the source domain latent-space statistics.

Matt Rounds

The awkward Edinburgh intellect who knows something about everything

What’s your background?

I’m a second year PhD student in Machine Learning and Cognitive Science at the University of Edinburgh, working on Bayesian models of human visual attention. My current focus is testing whether inattentional blindness (the phenomena where people miss very obvious things because they’re attending to something else) emerges naturally from the constraints on our cognitive system.

What did you work on during your internship?

I’ve been working with the internal R&D team on Explainable AI (XAI). It’s an important topic -whilst machine learning models are routinely applied to different problems here, clients often want to know why a particular prediction or output was produced. As the underlying model can be something of a black box, we need a way to generate clear explanations that are also true to the model. I’ve been looking specifically into deep learning-based methods to generate this simple explanation. Considering what constitutes a good explanation is central to the trustworthiness of this process -whilst the explanations are there to make the underlying model’s predictions make sense, we also need to be sure that those explanations themselves are sensible.

Mhamed Jabri

I have already completed two research internships. However, neither included exposure or contact with clients, leaving me with little understanding of how my research would be applied in practice. I was keen to discover what consulting looks like from a data science perspective when I discovered QuantumBlack on LinkedIn. I was immediately impressed with the profiles of QuantumBlack data scientists. These people possessed an excellent qualification in a specific research field, yet were also working across a hugely varied mix of interesting projects.

What did you work on during your internship?

I worked on a pharmaceuticals project. Days after joining, I was already traveling, meeting with clients and seeing my own work presented. Much of my activity involved digging into the client data and harnessing this to build models and offer solutions to their challenges.

What did you enjoy during your time at QuantumBlack?

QuantumBlack provided an incredibly warm and diverse environment. My 11 person team was made up of people from Greece, India, Croatia, Spain, Morocco, Libya and the UK. The surroundings were also tremendously supportive. I was greatly encouraged by the trust my colleagues placed in me and I did not once feel like ‘just an intern’ -I was a valued member of the team.

In 2019, we will run another internship programme across our Data Science, Data Engineering and Machine Learning teams. Applications will open in February 2019 and candidates can apply directly through our website. Shortlisted candidates will proceed to an interview process, involving:

  • CV review and an initial recruiter call
  • An online technical test
  • An onsite technical interview and the opportunity to meet and mix with the team

We look forward to welcoming new and talented interns through our doors next year!



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
QuantumBlack, AI by McKinsey

QuantumBlack, AI by McKinsey

An advanced analytics firm operating at the intersection of strategy, technology and design. www.quantumblack.com @quantumblack