CDS Industry Connections Lecture Series

Meeting the Challenge of Cracking Open Large-Scale AI with David Bau

Friday, April 18 | 2-3 PM
BU Computing & Data Sciences, 665 Comm Ave., Boston, Room 1646

About the Talk: What can humanity learn when we crack open black-box AI systems and look inside? In this talk, David will examine some recent insights from interpretable machine learning, asking how large-scale AI systems organize knowledge, Theory of Mind, and a "dual route model of induction" that uncovers a separation between token, concept, and functional representations during in-context learning.  After discussing recent work from the Bau lab as well as some inspiring results across the field, David will also discuss policy questions confronting machine learning including—How will we preserve human agency in a world of "superhuman AI?" Is interpretability essential for long-term progress in the field? And if so, what does it mean about how we should structure our industry today?

About the Speaker: David Bau is an assistant professor in the Khoury College of Computer Sciences at Northeastern University, based in Boston. Bau's research focuses on human-computer interaction and machine learning. Before joining Northeastern, he worked as a software engineer at Google, BEA, and Crossgain. He has been published in journals such as CVPR, NeurIPS, ICCV, ECCV, and SIGGRAPH. Outside of research, Bau enjoys astronomy and puzzle collecting.

How Wayfair is Transforming Customer Experiences with Data-Centric AI with Vinny DeGenova

Tuesday, April 8, 2025 | 4-5 PM
BU Computing & Data Sciences, 665 Comm Ave., Boston, Room 1646

About the Talk: Wayfair is transforming how customers discover home products through cutting-edge machine learning and data-driven automation. In this talk, we’ll dive into how AI— and partnering with Snorkel AI—enables us to extract high-quality product tags from supplier-provided labels and images, enriching our catalog and adapting dynamically to customer search behavior. These innovations drive tangible business impact, from increased add-to-cart rates to reduced cart abandonment and higher customer lifetime value.

This session isn’t just about algorithms—it’s an opportunity to engage with an industry leader working at the forefront of applied machine learning. Whether you're curious about real-world AI challenges, interested in the intersection of ML and e-commerce, or thinking about your own path in the field, bring your questions and let’s chat!

About the Speaker: Vinny is the leader of the Catalog Science organization at Wayfair, overseeing teams that apply machine learning and generative AI to understand the product catalog, extract insights from user-generated content, and develop multimodal foundation models for search, recommendations, and on-site experiences. His teams build visual search applications, resolve duplicate products, run competitive product matching algorithms, and predict future product performance.

With a background in Electrical Engineering (Boston University ‘14) and a Master’s in Computer Science, Vinny brings 10+ years of industry experience, including leading personalization ML teams and developing ML solutions in the defense industry.

Read the "3 Data Science Questions with Vinny DeGenova" blog post.

POSTPONED! Mastering Data-Centric AI: Unlocking the Full Potential of Your Computer Vision Datasets with the Open-Source Tool FiftyOne with Paula Ramos

March 21, 2025 | 1:30-2:30 PM

BU Computing & Data Sciences, 665 Commonwealth Ave., Boston, Room 1646

About the Talk: In computer vision, high-quality data is just as crucial as high-performing models. This 45-minute session will introduce data-centric AI workflows using FiftyOne, the open-source toolset for dataset visualization, embedding analysis, and model improvement.

You’ll learn how to:

• Visualize and explore complex datasets
• Analyze embeddings for deeper insights
• Improve model performance through advanced data curation
• Leverage plugins and integrations for workflow automation

About the Speaker: Paula is a results-oriented AI professional, expert in technical solution architecture, community evangelism and driving adoption of Conventional AI, GenAI and IoT. As an experienced solution architect, Paula loves collaborating with cross-functional teams, defining clear product requirements based on end-user needs, and ensuring seamless integration of new AI technologies, including GenAI, into complex systems, both at the edge and in cloud-like environments with microservices. As an AI Evangelist, Paula excels at creating and fostering dynamic developer communities, fostering strong relationships with ISVs and developers, and leading the creation of scalable meetups and user groups.

Innovation through Data @ MassMutual with Adam Fox

October 15, 2024 | 4:00-5:00 PM

BU Computing & Data Sciences, 665 Commonwealth Ave., BU Spark! 2nd Floor Community Space

About the Talk: Over the past decade MassMutual has leveraged novel data and algorithms to drive innovation across the client lifecycle, from marketing to underwriting, product design, and investment management. In this talk we will begin with an overview of Insurance and Financial Services and explore how data science has been leveraged to improve decision making and drive efficiency. We will then transition to discuss emerging trends in Generative AI and Wellness.

About the Speaker: Adam is the Head of Distribution Technology & Data Science at MassMutual and leads MassMutual’s Distribution Technology and Data Science organization, where his teams are responsible for providing holistic technology, data, and digital support for MassMutual distribution as well as enterprise data science enablement. Prior to joining MassMutual, Adam served on the mathematics faculty at the Georgia Institute of Technology and Western New England University where his research focused on forecasting the long-term outcomes of complex systems in nature, the economy, and human behavior. Learn more about Fox in the "3 Questions with Adam Fox" news post.

Bringing 21st Century Solutions to Heating and Cooling Buildings with BlocPower with Craig Altemose

April 30, 2024 | 4:00-5:00 PM

BU Computing & Data Sciences, 665 Commonwealth Ave., Room 1646, Boston

About the Talk: BlocPower is working to electrify every building in America, leveraging a model of horizontal integration that includes technology tools to identify and prioritize candidate buildings to electrify, in-house engineering and project management expertise, program management to assist cities, utilities, and states in achieving their building electrification goals, and workforce development programs to close the labor gap in the building electrification industry. Craig Altemose will share about BlocPower's model, how it partners with universities, and what it looks for from academic partnerships.

About the Speaker: Craig Altemose is the Director of Strategic Partnerships at BlocPower, a black-owned climate technology company working to make buildings smarter, greener, and healthier through upgrading heating and cooling systems to 100% highly efficient electric alternatives. Prior to joining BlocPower, Craig served as the founding Executive Director of Better Future Project, a climate nonprofit based in Massachusetts. Craig holds a J.D. from Harvard Law School, an M.P.P. from Harvard Kennedy School, a B.A. from Eckerd College, and a Certificate in Nonprofit Leadership and Management from BU's Questrom School of Business. He lives with his wife and young sons in Maynard, MA. He previously served on the Massachusetts Climate Protection and Green Economy Advisory Committee.

Human-AI Collaboration & Adaptive Processes in Industry with Enrico Santus

April 2, 2024 | 4:00-5:00 PM

BU Computing & Data Sciences, 665 Commonwealth Ave., Room 1646, Boston

About the Talk: According to Paul R. Daugherty and H. James Wilson, authors of “Human+Machine: Reimagining Work in the Age of AI,” organizations are now going through a third wave of business transformation. The first wave started when Henry Ford deconstructed the manufacture of automobiles, standardizing processes. In the ’70s, the second wave targeted the automation of processes through information technology. The third wave, which started only recently, focuses on adaptive processes, where the optimization does not target the maximization of the efficiency of individual steps, but rather looks at the outcome more holistically, and the business impact.

While many companies are using AI to automate processes, those that mainly deploy this technology to displace employees will see only short-term productivity gains. In a Harvard Business Review article Daugherty and Wilson shared that among 1,500 companies significant improvements are achieved by firms where humans and machines work together. Thanks to the introduction of Large Language Models (LLMs), AI is becoming more accessible to humans, who can use natural language to achieve goals that would have before required code. Moreover, LLMs broaden the scope of AI, covering not only labor-intensive tasks, but also strategic and creative ones, such as design, marketing, customer service, and sales.

Through collaborative intelligence, humans and AI actively complement and enhance one another’s strengths: the leadership, teamwork, creativity, and social skills of the former, and the speed, scalability, and quantitative capabilities of the latter. In this talk, Enrico Santus will discuss how - also thanks to LLMs - the human-AI collaboration can be designed and optimized.

About the Speaker: Enrico Santus, a Sardinian native, embarked on an academic journey that led him to a doctorate in Computational Linguistics, supported by a prestigious fellowship. His pioneering research in deep learning took place at MIT’s CSAIL, and he later joined Bayer before becoming the Head of Human Computation at Bloomberg’s CTO office in New York, where he envisions the future of active learning and human-in-the-loop annotation. His work spans diverse fields, from fake news detection and healthcare data extraction to pharmaceutical research, and has earned recognition from institutions like the White House and the American Congress.

Learning from Failure: How a Global Outage Can Improve System Design with Jeromy Carriere

February 20, 2024 | 4:00-5:00 PM

BU Computing & Data Sciences, 665 Commonwealth Ave., 17th floor, Boston

About the Talk: In March 2023, Datadog experienced a multi-hour global outage of its services, impacting almost every customer. This talk will discuss the outage timeline, its root cause, and the factors that contributed to its breadth and duration. But more importantly, we'll discuss what we learned from the outage, and how it informs ongoing improvements to Datadog's systems and principles of design.

About the Speaker: Jeromy is Senior Vice President of Engineering at Datadog, where he leads engineering for all of Datadog’s product areas including Infrastructure Monitoring, Log Management, Application Performance Monitoring, Security, Service Management, and User Experience Monitoring. Prior to Datadog, Jeromy was an engineering director at Facebook, leading a team building a service management platform for Facebook's private cloud. Before that, he was at Google, where his teams built the Google Stackdriver product suite for monitoring, logging, and tracing as part of the Google Cloud Platform. Jeromy previously played architect and management roles at eBay, Yahoo!, Vistaprint, Fidelity Investments, Microsoft, and AOL, and co-founded Kinitos and Quack.com.

Preparing for Industry with Marina Hatsopoulos

Wed., November 29, 2023 | 4:00-5:00 PM

BU Computing & Data Sciences, 665 Commonwealth Ave., 17th floor, Boston

About the Talk: Entering the workforce is a daunting transition which requires a shift in mindset. Important decisions which will impact your professional trajectory can be made easier after some introspection about what makes you unique. With most net new jobs in the U.S. created by startups less than five years old, you may want to consider the benefits of working at a startup as compared to an established company. Once in the workplace, conflict will be unavoidable, and how you approach these scenarios can impact your career growth. This talk is intended to help guide your thinking as you transition to a new phase of life without the framework of school.

About the Speaker: Marina Hatsopoulos is Board Chair of Levitronix Technologies, the worldwide leader in magnetically-levitated bearingless motor technology, and President of Hellenic Innovation Network, which serves as a bridge for startups expanding from Greece to the U.S. She is on the MIT Corporation Visiting Committee for the Department of Mechanical Engineering. She is on the Advisory Board of the Nantucket Conference, MIT Enterprise Forum Greece, Eurobank’s EGG accelerator, Mindspace Entrepreneurship Program and MIT ReACT. She was Founding CEO of Z Corporation, an early leader in 3D printing out of MIT, and has served on numerous corporate boards, both public and private. She was a Director of Cynosure Inc., a $400 million leader in the laser aesthetics market; GSI Group, a $300 million supplier of laser-based equipment; and Contex Holding, a $100 million leading manufacturer of large-format scanners.

She speaks regularly on topics related to entrepreneurship at MIT Bootcamps and all over the world, and profiles about her have run in Fortune, Forbes, Fast Company, Boston Business Journal, Boston Magazine, and Technology Review. Her business writing has appeared in Venturebeat, The Observer, CEO World Magazine, Design News, Time Compression Technologies, and Jumpstart: Launching Your Business Venture, Profitably and Successfully (Aspatore, 2003). She was Speaker, Tedx. She has also written prize-winning essays and short stories which have been published in Antioch Review, Missouri Review, Bellevue Literary, and other literary journals. She holds B.A. degrees from Brown University in Math and in Music, and an M.S. from the Massachusetts Institute of Technology in Mechanical Engineering.

Read "3 Questions with Marina Hatsopoulos" 

The Evolution of Media Technology for the Online World with Jeremy Doig

Tues., November 7, 2023 | 4:00-5:00 PM

BU Computing & Data Sciences, 665 Commonwealth Ave., 17th floor, Boston

About the Talk: In this Industry Connections talk, Jeremy will discuss the origins of and motivation for Google's significant investment in free, open-source media technologies and how the industry needed to change. Looking forward, he will cover new areas of development and opportunities for engineers entering the field.

About the Speaker: Jeremy Doig's history spans large-scale systems and multimedia at the BBC, Apple, Microsoft, Google, and Disney—in that time, he has pioneered new standards for online media, spanning novel compression approaches for audio and video, streaming protocols for real-time and on-demand delivery, and spatial experiences.

Read "3 Questions with Jeremy Doig"