S1E1: Meta's senior machine learning engineer Katerina Iliakoupoulou on leaving your dream job and the future of ML
Jul 29, 2024
auto_awesome
Katerina Iliakoupoulou is a senior machine learning engineer at Meta, previously with the New York Times. She discusses her surprising decision to leave the journalism world for tech, emphasizing the necessity of deep domain knowledge for engineers. The conversation shifts to building impactful ML solutions, particularly for Facebook Reels, highlighting the importance of teamwork and iteration. Katerina also shares insights on the future of recommendation systems and the evolving role of machine learning in enhancing user experiences.
Deep domain knowledge significantly enhances an engineer's effectiveness by fostering innovative solutions and enabling a nuanced approach to problem-solving.
The rapid evolution of machine learning and user engagement will transform recommendation systems into personalized, interactive experiences that adapt to individual feedback.
Deep dives
The Importance of Domain Knowledge in Engineering
Deep domain knowledge is crucial for engineers, significantly enhancing their effectiveness in their roles. A strong foundation in both engineering skills and an understanding of the specific industry leads to more passionate and competent work. For example, the speaker's journey from journalism to machine learning illustrates how combining expertise can foster innovative solutions. Passionate involvement in one's domain enables engineers to tackle problems with a more nuanced perspective, ultimately driving better results.
Transition from News to Machine Learning at Meta
The speaker detailed her transition from the New York Times to Meta, emphasizing the motivation to work with cutting-edge technologies in machine learning. Initially starting as an intern at the New York Times, she advanced to a tech lead role, focusing on personalization and messaging platforms. Though it was a challenge to shift from news to a broader scope of content discovery, the move allowed her to engage with a larger audience and tackle significant information consumption issues. This change highlights the importance of adaptability in fostering new opportunities and growing as a professional.
Adapting to a Rapidly Changing Environment
Working in a large organization like Meta presents unique challenges, especially regarding prioritization and staying informed in a fast-paced atmosphere. Engineers must be proactive in aligning priorities and keeping track of ongoing projects, which requires efficient communication and collaboration. Regular check-ins and utilizing internal platforms for updates are essential strategies for remaining informed and engaged with team goals. This dynamic environment emphasizes the importance of flexibility and continual adaptation in the engineering process.
The Future of Recommendations and Machine Learning
The speaker predicts a transformative evolution in how users engage with recommendations through interactive AI systems that adapt to personal feedback. As audiences increasingly consume content across various platforms, personalization will become more responsive, allowing users to directly input their preferences and desires. This trend highlights the need for companies to innovate and make their content easily accessible beyond traditional media formats. Ultimately, the interplay of machine learning advancements and user engagement will shape the future of content consumption.
Katerina Iliakopoulou, a senior machine learning engineer at Meta, joins the show to talk about leaving her dream job at the New York Times, the importance of deep domain knowledge as an engineer, and what’s around the corner in machine learning.
02:10 Why domain knowledge is vital for engineers
04:45 Leaving the New York Times for Meta
08:45 Choosing the IC track and what makes a great manager
14:44 The art of prioritization and building impactful solutions
18:00 PriorityZero: Scaling ML models for Facebook Reels