Atlassian Career Growth & Development

Give us a snapshot of what you do to stay on top of your engineering knowledge and skills. 

Learning schedule: As a machine learning engineer at Atlassian, I leverage the flexibility of remote work to dedicate my mornings to focused learning. Living in Seattle with collaborators in Sydney, I am able to keep my mornings relatively meeting-free to have precious focus hours. 

Literature review: Social networking sites can often overflow with content where users share their opinions on emerging technologies. I find it significantly more time-efficient to delve into the original research papers authored by the inventors. Much like I prefer reading a comprehensive book on a subject that truly interests me, rather than skimming through blogs or watching YouTube videos. 

Building and reproducing: From an engineering perspective, I only feel I understand a system when I can rebuild them, so I spare my personal time to reproduce systems whenever possible. For instance, in my exploration of large language models, I loved Andrej Karpathy’s amazing guidance in building NanoGPT. I followed the same footprint to write a GPT from scratch and use them as a test ground to apply other techniques on top. This way, I understand every line in my Lego blocks and can reorganize them as I learn new concepts. 

Comparative analysis: We consistently evaluate our implementations, including agent orchestration, against open-source alternatives to monitor our progress and identify areas for improvement.

What are your go-to resources for keeping your engineering skills sharp? 

In engineering, the process of developing new products offers the exciting opportunity to create solutions from the ground up. With the emergence of numerous machine learning models, frameworks and advancements in generative AI, we find ourselves presented with a wealth of possibilities. However, it can be all too easy to remain focused on high-level technological trends. I make a conscious effort to stay grounded in practical applications. Engaging in pair programming and participating in extensive code reviews are essential practices for me. When time constraints prevent me from meticulously building a new feature, I prioritize code cleanup. This practice not only helps me stay current with recent code changes but also deepens my understanding of the evolving codebase. Earlier in my career, I made the mistake of learning just enough to perform my job, rather than developing a deeper understanding of the technologies I was working with. Now, I consciously focus on the compounding benefits of building a solid foundation.

What does the learning culture look like at your company or on your engineering team? 

Collaborative learning proves to be more effective, which is why we have successfully implemented several group learning programs. This collaborative environment fosters a vibrant and dynamic learning culture at Atlassian. 

AI School: As Generative AI continues to gain traction, an increasing number of software engineers are focusing their efforts on this innovative field. To support this trend, we have established an AI school that offers foundational courses in machine learning (ML 101) as well as training on Atlassian-specific platforms and tools. This initiative is designed to reach a wider audience of software engineers, enhancing their skills and knowledge in AI. 

ML Reading Group: Our biweekly ML reading group encourages vibrant discussions on the latest research papers, attracting enthusiastic participation across the orgs who often sign up months ahead of time. 

AI Summits: We organize cross-organizational AI summits where each team presents their experiences in developing AI products. These sessions cover their methodologies, engineering challenges, product insights and key learnings.