Principal Applied Scientist, Machine Learning

Key Skills

Python Programming, DevOps, AWS, Scala Programming, Machine learning techniques, Docker, Kubernetes-K8s, Continuous Integration & Continuous Delivery - CI/CD,

Job Description

We are the Data Team. A collective of specialists driven by our shared mission to help establish Xero as the most insightful and trusted small business platform. Our backgrounds and work are diverse—we work on data platforms, data modelling, application development, applied research, ethics, literacy, and strategy, to name a few. At the heart of our mission is helping our colleagues and customers get value from data responsibly, on robust data platforms, with robust methods.

Applied Scientists work in cross-functional teams (engineers, designers, product managers, etc) to create AI-powered experiences. We apply machine learning to develop products that reduce toil and provide insight, allowing small businesses to focus on what matters.

There’s a myriad of job titles out there and the English language can only express so much in a couple of words. So role title aside, what we’re looking for is someone to:

About the role:

• Provide leadership to cross-functional teams that build web-scale ML products
• Work with your peers and the Data Leadership Team to define the roadmap for Applied Sciences
• Identify cross-cutting opportunities to create frameworks / tooling and improve data quality
• Work with other scientists to figure out the best practices for reproducible and robust science, before documenting and evangelising them
• Keep up-to-date with the latest developments in ML, identifying opportunities for Xero to improve #beautiful products.
• Be comfortable with pushing the frontiers of ML in the pursuit of delivering products, but only where absolutely necessary
• Teach colleagues around Xero about how use data the right way
• Make sure our work is always in the best interests of our customers
• Align multiple programs of work so that they benefit each other
• Identify new ML product opportunities and work with product teams to determine if they’re worth pursuing
• Create the culture that makes people love working with us
• Mentor, manage, and develop other scientists and colleagues

About you:

• You have extensive experience delivering web-scale production machine learning systems
• You rejoice in identifying tough problems that can be solved with scientific thinking
• You understand that solving worthwhile problems means covering all areas of the data lifecycle - finding it, understanding it, experimenting with it, despairing over it, fixing it, …
• You’re comfortable reading ML research papers, keep abreast of new work on arXiv.org and have on occasion spent your time tinkering around with that interesting new framework you read about recently ... but when it comes down to it, you want to solve real business problems in the simplest and most robust way possible
• You’re comfortable working with experts from all corners of the company to deeply understand the customer problem before getting your hands dirty with all that data and code
• Experience as a hands-on practitioner building productionized machine learning pipelines which touch real human end users
• You have a solid grasp of statistics and you have wrestled with the challenges inherent in actually measuring the impact of your machine learning pipelines in the wild
• You’ve learned through experience that it’s important to write testable, repeatable code so you can debug it when something goes bump in the night
• You know your way around the ‘nix command line, ssh your way happily around your multiple running AWS instances and are a competent programmer in Python, Scala or similar
• You’ve mentored and grown teams of applied scientists, teaching the skills of scientific reasoning, collaboration, and effective program delivery.

What you'll bring:

• Ability to translate between the business and machine learning domains
• Sound software engineering skills in Python, Scala, or similar
• Expert level hands on practitioner skills in one or more of: time series forecasting, natural language processing, classical machine learning, deep learning
• Ability to grasp and teach the mathematical concepts that underpin the domains outlined above
• Comfortable with version control and command line
• Strong communication skills and the ability to tailor a message for peers, senior stakeholders, and junior team members alike
• Ability to mentor and grow junior scientists