Research Scientist, AI/ML (PhD)

Ads is the largest revenue generator at Meta and Ads Quality represents around 20% of total revenues which are used to generate long term ads and organic engagement.Core Ads Quality is a unique team jointly optimizing for both quality and revenue, aiming at making this investment more revenue / quality trade-off efficient and generate long term revenue growth through user learning. Among others, Core Ads Quality focuses on:- Finding the right trade-off between short and long term revenues- Standardising and optimise quality treatment of ads across surfaces and page types- Understanding user behaviour with respect to ads quality- Building a solid infrastructure around signals, labels and quality metricsWe work at the intersection of Ads, Machine Learning and User Behaviour understanding. The nature of our work is very analytical, with a solid collaboration with our Data Scientist and a heavy focus on not only understand “what” but also “why”. Despite having been created a couple of years ago, the Ads Quality space at Meta is still nascent and full of unexploited opportunities. The organization is further structured into the following teams/sub-pillars:- Integrity & Efficiency: Proactively cover long-term revenue risks from advertiser friction while supporting XI with delivery expertise.- Ads Conversion Familiarity: Accelerate Non-Purchaser (NP) -> Purchaser (P) transition by increasing familiarity of ads for users who don't interact with ads frequently- Post-Click Quality: Stop Purchaser (P) - >Non purchaser (NP) user conversions from bad purchase experiences.- Modelling: Enhance quality and drive long-term revenue growth through modelling.- Quality Science: Build the foundational end to end understanding for Funnel quality signals to ensure its the efficiency, health and coverage.The team has consistently hit their goals and delivered XXXM$ in incremental long term revenue for Meta while ensuring high ads quality.

Responsibilities

• Apply relevant AI and machine learning techniques to build Ads auction/user behaviour treatments
• Develop novel, accurate AI algorithms and advanced systems for large scale applications
• Work towards long-term ambitious production goals, while identifying intermediate milestones
• Directly contribute to experiments, including designing experimental details, developing reusable code, running evaluations, and organizing results
• Work with large data, and contribute to development of large scale foundation models
• Design methods, tools, and infrastructure to push forward the state of the art in AI

Qualifications

• Currently has or is in the process of obtaining a PhD in Artificial Intelligence (AI), Computer Science, or a related technical field
• Experience developing machine learning algorithms or machine learning infrastructure in Python, PyTorch, and/or C/C++
• Experience in training, fine-tuning, and/or experimenting with foundation models beyond black-box use Experience in Reinforcement Learning, GenAI, Large Language Models, etc
• Experience in Ads, especially in auction theory and implementation (bidding, budgeting, targeting)
• Experience in User Behaviour modellling, Long-term Value optimization or Causal Learning