Mastering Personalization

& Recommendation Technology

Your chance to engage live with a leading expert on personalized machine learning, Professor Julian McAuley, and a cohort of driven industry professionals

5 x 2 hr sessions (+ recordings of each live session)

8am PST; June 6, 9, 13, 16, 21

$750 per seat (expense through L&D)

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.

Your referral makes you eligible for a $50 Amazon Gift Card!

Designed for data, engineering and product specialists looking to implement

Personalization Systems

ML Engineers

tasked with instrumenting and deploying recommendation technologies

Deployment considerations and common pitfalls

Practical tradeoffs between recommendation approaches

Strategies for measurement and evaluation of recommender system performance

Data Product Managers

running product teams developing recommender systems

Qualitative concerns of recommendation evaluation (diversity, serendipity, etc.)

Ethical concerns in recommendation and personalized machine learning

Emerging technologies and recommendation paradigms

Data Scientists

tasked with developing and implementing recommender systems and personalized predictive models

Build predictive systems involving user data

Learn how to model different types of interaction data (clicks, likes/dislikes, purchases)

Strategies for model training and optimization

Meet your

Instructor Prof. Julian McAuley

Julian McAuley is a Professor of Computer Science at the University of California, San Diego, and is currently an Amazon Scholar. His lab's research primarily focuses on applications of machine learning to problems involving human interactions, ranging from personalized healthcare to fashion design.

Julian’s lab regularly collaborates with industry partners to develop recommender systems, including Microsoft, Facebook (Meta), Pinterest, Salesforce, Nvidia, Samsung, Adobe, Toyota and many others. His lab is also supported by an NSF CAREER Award.

Julian’s papers have over 15,500 citations and has also authored the book ‘Personalized Machine Learning’ which was recently published by Cambridge University Press.

About Julian’s

Live Cohort

Every day we interact with models that make predictions or recommendations for our entertainment, social connections, purchases, or health. Such models leverage several modalities of data, from sequences of clicks to text, images, and social interactions. This course will introduce common principles and methods that underpin the design of personalized recommendation technology.

Learn how to design, implement, and evaluate recommender systems, including case-studies of recommender systems from Amazon, Netflix and many more. Learners will understand the main modeling techniques involved, ranging from simple neighborhood-based models (“people who bought X also bought Y”) to complex models based on deep learning. We’ll also discuss deployment issues and ethical concerns, including issues related to content diversity, filter bubbles, and qualitative considerations.

After taking this course, you will be able to:

Understand the design principles behind recommendation technology, and how it differs from other types of machine learning

Build recommender systems involving different modalities of data, including sequential, textual and visual data

Evaluate and compare recommender systems, including quantitative and qualitative metrics (relevance, novelty, serendipity, etc.)

Navigate the common pitfalls of deploying recommendation technology, including issues of ethics, fairness, and content diversity

Session 1 - Introduction and “classical” recommender systems

Monday - 6th June
8am PST / 4pm GMT

Heuristic techniques to uncover patterns of similarity among users or items

Methods that drive simple item-to-item recommenders (“people who viewed X also viewed Y” etc.)

The relationship between simple heuristic methods and ML-based methods

Common techniques and pitfalls behind evaluation of recommender systems

Session 2 - ML and “model based” approaches to recommendation

Thursday - 9th June
8am PST / 4pm GMT

The relationship between recommendation and other types of ML (including regression, classification, and dimensionality reduction)

Strategies to predict different types of interaction data (including ratings, clicks, purchases etc)

Recent trends to model interaction data based on deep learning

Session 3 - Content and structure in recommender systems

Monday - 13th June
8am PST / 4pm GMT

How to apply recommenders in scenarios outside of normal (“e-commerce-like”) settings

Design principles for adapting personalized models to different types of data

Related personalization scenarios involving text, conversation, fashion recommendation

Session 4 - Modeling sequential and temporally-evolving user data

Thursday - 16th June
8am PST / 4pm GMT

Understand the critical role that temporal dynamics play in capturing evolving user behavior

Techniques to incorporate different types of temporal trends (including seasonal dynamics, long-term trends, and short-term sequential context)

Case studies (e.g. recommendation on Netflix) that highlight the overall design approaches to modeling temporal dynamics

Session 5 - Ethics and fairness in recommender systems

Tuesday - 21st June
8am PST / 4pm GMT

Common pitfalls of recommendation when applied naively, including filter bubbles, fairness, and diversity issues

Case studies from Youtube, Facebook, Netflix, and Spotify (among others) that highlight these issues

Algorithmic strategies to correct such undesirable outcomes

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.

Based on collaborations with