Example projects

The D3 group is a central team that supports all areas of the enterprise. At any given time, there will be tens of R&D projects in progress. Here are just a few examples:

• Kernel smoothing of spatial data for weather risk modelling
• Gradient boosted trees and logistic regression for predicting probability of customer defection
• Supervised and unsupervised graph mining for fraud detection
• IoT: Unlocking the value in data generated from connected vehicles
• Modelling business objectives as high dimension constrained numerical optimisation problems
• Deep neural networks for classification of photographic claim evidence
• Natural language processing for customer sentiment analysis
• High performance distributed data pipelines with Apache Spark
• Generalised linear models for predicting claim frequency and severity
• Bayesian methods for credibility and change point detection

The environment

We are a research group: we work on long-term projects; we read scientific papers and textbooks; we talk about ideas; we explore new programming languages; we huddle around whiteboards; we try out the latest machine learning algorithms; we think about what the company should look like next year and in five years.

And we get things done: we develop the next generation of products; we crank out code; we do scrum and agile; GitHub and Jenkins; code reviews and sprint reviews. We invest heavily in cutting-edge software and hardware and encourage experimentation: we count CPU cores by the hundred and RAM by the terabyte. We use and contribute to open source software and work mainly on Linux.

Your role in the team

Each year we invite a small group of individuals to work with us for a few months. We look for those with a passion for finding great solutions to complex problems, and enthusiasm for quickly learning whatever techniques and technology are right for the job.

Successful candidates will work from our Belfast office during the summer of 2021 among more than 30 experienced data scientists and engineers and will work with a team that spans multiple locations. They’ll learn about how data science techniques are used to solve business problems, then apply this knowledge on a challenging research project which will be of real value to Allstate.

• Solve difficult problems
• Use mathematics
• Write code
• Work with lots of data

Those who successfully complete their 12-week internship with Allstate Northern Ireland may be made offers of permanent employment with Allstate. These positions will be conditional and commence post-graduation and will be subject to the student securing a first-class classification in their BSc degree or a Distinction or equivalent in their Master’s Degree or PHD.

So, what are the essential criteria to apply?

• Must be a UK/RoI/EU citizen or possess a current UK visa and eligible to take up full time, permanent employment. EU candidates must also demonstrate they are eligible to take up UK employment post-Brexit.
• Must be eligible to travel to the US
• First class degree (attained or expected) in a highly quantitative field such as mathematics, physics or computer science
• In final year of a Bachelor’s degree OR
• In final or penultimate year of a Masterʼs degree OR
• In final or penultimate year of a PhD program

We also have some desirable criteria

• Working towards a Master’s degree or PhD in a highly quantitative field
• Final year student
• Proven experience in general purpose programming such as Python, Scala, Java or C/C++
• Experience in any of the following:
• Machine learning or statistics to solve real-world problems
• Statistical software and libraries such as R, scikit-learn or MLlib
• Distributed computing or big data frameworks such as Spark, Mesos or Hadoop
• High performance and parallel computing such as openmp, MPI or GPU programming
• High dimension discrete or continuous numerical optimisation

What we offer

• A competitive salary
• Extensive and thorough in-house training
• A globally collaborative working environment
• Invaluable work experience
• Permanent offers may be available upon successful completion of the internship