A Deep Dive into the Life of a DeepMind Software Engineer
In this edition of ‘Five minutes with’, we sit down with Kevin Millikin, a talented software engineer on the DevTools team at DeepMind. Currently, he is attending PyCon US, the largest annual gathering for Python programmers, to deliver a presentation. Let’s take a closer look at Kevin’s role and his insights on the development of AI.
Kevin’s Work at DeepMind
Kevin specializes in creating tailor-made software tools for developers at DeepMind. For instance, his team is currently focused on developing a web-based editor to facilitate coding in Python, a widely-used language among DeepMind’s engineers. These tools are essential in supporting their work and leveraging the Google infrastructure effectively, enabling them to tackle significant problems that impact their teams.
A Typical Day in Kevin’s Life
Kevin begins his day at the stunning London campus of DeepMind. They follow a hybrid work model, spending Monday through Wednesday in the office and Thursday and Friday working remotely. Kevin particularly enjoys the face-to-face interactions with his colleagues, as it fosters a collaborative work environment.
On Thursdays and Fridays, Kevin works from his home office, which also happens to be his music room. As a musician, Kevin plays several instruments, including the bass guitar, baritone horn, and tenor saxophone. He found that playing music during the remote work period of the pandemic provided him with a different creative energy. It allowed him to reflect on the problems he was solving and approach them from different angles, leading to innovative solutions.
Kevin’s Presence at PyCon US
At PyCon US, Kevin is excited to present his talk on ‘Beyond Subtyping,’ a crucial Python feature. His session provides insights into the disagreements that arise among the tools implementing subtyping. As a Python designer, one might assume these issues are resolved, but there are still foundational points about the language that lack consensus.
Kevin actively participates in the typing working group, collaborating with professionals from companies like Microsoft, Facebook, and Google. Their goal is to evolve Python to meet the needs of their respective user bases. Through this collaboration, they have discovered common problems and shared aspirations. The group strives to design tools that can benefit everyone, necessitating a highly collaborative approach.
Kevin’s Areas of Excitement
Kevin is thrilled about the opportunity to meet face-to-face with colleagues he has worked with remotely for the past few years, all of whom are part of the Python language community. He aims to expand their network and foster inclusivity for external contributors. Kevin believes that the work done in this field tends to be confined to a closed group, and he sees great potential in opening it up to a broader audience.
The Future of Language and AI
While Python continues to evolve with the addition of new features, it often becomes challenging to ensure coherence among these features. Kevin advocates for taking a step back and defining principles for the evolution of this programming language. He believes it would be beneficial to document these principles as a roadmap for the next 2-5 years, enabling more thoughtful and future-oriented changes to the language. This foresight ensures the development of tools that will accelerate AI research and meet the needs of the ever-evolving field.
For more information on engineering at DeepMind and available positions, please check our website.