Research Scientist TLM - Statistical Learning and Experimentation

Key Skills

Apache Hadoop, C Programming, C++, Data science techniques, Java Programming, Machine learning techniques, MapReduce, Python Programming, PyTorch, R Programming, Scala Programming, SQL, TensorFlow

Job Description

Our Statistics & Privacy team focuses on establishing step-change improvements to advance our long-term business growth. The team's expertise spans many domains including statistics, machine learning, economics, and cryptography. We develop methodologies, design and prototype solutions, and partner with our engineering colleagues to launch these solutions such that 3 billion people and 200 million businesses on our platform can benefit. Examples include internal and external experimentation platforms, ads delivery optimization for efficient causal outcomes, and the application of privacy-enhancing technologies across our ads product stack. Facebook seeks an industry-experienced Research Science leader to join Statistics & Privacy. This managerial role is designed specifically for a Research Science leader to remain hands-on for technical and scientific development, while also directly supporting a small team of research scientists, and leading on a small set of org-level capacity-building responsibilities. Cutting-edge applied research necessitates identification and evaluation of new research opportunities, and development of scientifically rigorous systems. Research scientists drive forward prototypes and methodologies, and build consensus with engineering stakeholders as to the scientific and strategic merit of our initiatives. This role will focus on ranking, delivery, and experimentation systems that power our advertising and commerce capabilities.

Research Scientist TLM — Statistical Learning and Experimentation Responsibilities

• Hire and lead highly skilled research scientists, providing mentorship and career guidance to members of the team
• Identify specific, high-impact opportunities to build new and improve existing advertising and commerce solutions based on deep learning (retrieval, ranking, recommender systems), experimentation and inference systems
• Design and prototype new algorithms, optimization methods, and system architectures with hands-on implementation and development
• Shepherd the team’s research and development projects, ensuring high-quality communication across diverse leadership audiences, and taking responsibility for the team’s long-term product impact
• Build cross-functional relationships to shape long-term roadmaps with a balance of scientific rigor and strategic considerations
• Learn new tools, systems and languages quickly as required by the particular project you are working on
• Manage the development of presentations, patent applications, white papers and publications of research results in top-tier journals and at leading conferences, in a manner that is broadly appealing and accessible beyond a core scientific audience
• Lead author of publications at top-tier peer-reviewed journals or conferences (e.g. NeurIPS, ICML, JSM, KDD)

Minimum Qualifications

• PhD in a quantitative field such as statistics, computer science, physics, or computational social science and 4+ years of relevant experience, or MS/MA degree with 8+ years of relevant experience
• Expertise in statistics (particularly causal inference and Bayesian inference) or deep learning (particularly ranking, retrieval and recommender systems), including addressing challenges from incomplete, unrepresentative, and mislabeled data
• Proficiency designing and implementing analytical and/or algorithmic solutions, tailored to particular business needs and tested on large data sets
• Proficiency in Python, Julia, or R, including leveraging large (TB) datasets
• Proficiency in communicating analyses and establishing confidence among audiences who do not share your disciplinary background or training
• Track record of product or systems innovation, and strategic focus

Preferred Qualifications

• 4+ years of experience in leading and developing highly technical and/or applied research teams, including performance management and recruiting