What's your job about?
The immense potential of deep learning is rapidly being unlocked across multiple domains. The work that Dolby does in the audio and video space is no exception and there is intense effort underway to develop new algorithms that leverage the strengths of this technology. The potential of deep learning is matched by the engineering challenges it presents in going from a research prototype to a product. My job revolves around trying to overcome these challenges through the development and maintenance of shared tooling that expedites the development of these systems. Achieving this leverages a combination of computer science, an understanding of deep learning, and an ongoing commitment to understanding how the teams within the organisation work to develop these systems. A typical project will involve canvassing teams to understand how they are currently approaching a particular component of their workflow. With this understanding, I work with my manager and colleagues to develop an approach that meets their needs and improves upon their current approach. From there, I will begin building a prototype that can meet their immediate needs as soon as possible. This prototype will then be taken back to them, and we will work together to integrate it into their workflow. Their feedback while then be used to guide the tool’s continued development, and we will continue to find other teams facing the same problem and see if this tools meets their needs. If it does, we integrate, if it doesn’t, we work to understand why and adapt it.
What's your background?
I am a born and bred Canberran, having spent almost all of my life living in the bush capital. I went through the public education system with a heavy STEM focus before going to the Australian National University to study engineering/chemistry. I decided to pivot entirely to chemistry, but along the way stumbled into computer science, and in particular machine learning. This culminated in an Honours year spent researching how we can apply machine learning to chemistry problems. I enjoyed this so much that I moved onto a PhD working on this challenge. The PhD presented fascinating problems, and amazing opportunities – I was fortunate to spend 6 months at MIT in Boston working in the Green group. Towards the end of the PhD, I realised that the computer science problems were what I truly loved and that I wanted to step full time into those. I began preparing heavily for technical interviews and applied for a data analyst role at Dolby. Critically, I reached out to a recruiter, explained my position and background, and said that I was open to other roles. The recruiter got back in touch and said we have this other role that we think would perfectly match your background in scientific research, ML, and computer science. I applied, and, as they say, the rest is history. I’ve been working in the role now for about 6 months and I feel that I get to bring the full weight of my past experiences to the role, something I am immensely grateful for.
Could someone with a different background do your job?
Because this role has multiple different responsibilities, I certainly think someone with a different background could do it, but how they approached the role would likely change. Someone from a pure computer science background would likely spend more time coding the solutions, while someone from a different background might instead lean more heavily towards facilitating the teams themselves developing these tools and then simply manage the projects between them.
What's the coolest thing about your job?
The best part of my job is that I get such a high level, detailed understanding of the processes within the organisation, despite not being in any managerial position. I then get a fairly open slate to start trying to help those processes along. The most edifying part of the process is when someone contacts me out of blue and says “I saw this tool and our team wants to use it, can you show us how?”.
What are the limitations of your job?
Sitting between a large number of teams in different countries, one of the big limitations can be meeting times outside of work hours. The other limitation is that the responsibility placed on you increases as more people use your tools and infrastructure. While it sounds like a catch 22, I think many jobs have this feature – ideally, if you are doing something right, more people will want to use it. That’s the goal after all!
3 pieces of advice for your university-self?