I am a Research Scientist Manager at Meta. My research is focused on improving the quality of ads using advanced machine learning. I'm motivated by solving real-world problems where the requirements do not fit into a classic machine learning scenario (e.g., supervised, contextual bandits, or reinforcement learning). A few examples of interesting twists I've worked on include optimizing for a constrained objective, learning classifiers that are robust to adversarial attacks, managing risk, and learning from delayed signals.
Transfer learning, reinforcement learning, representation learning, online advertising, recommender systems