Jasmeet is the VP, Decision Science Analytics at Truist, and is also the current winner of the Rising Star in Data and Analytics award, at the CDAO Conference in Boston.

This episode is an opportunity to look behind the curtain, getting a detailed view of the business and volunteer profile of a top award winner in our industry. Also, near the end of the episode, Jasmeet describes an approach for ensuring that an analytics team is viewed by their company as a profit center, not as a cost center.

Michael holds a PhD in Economics from the University of Pennsylvania, and is a Senior Fellow at Wharton (Artificial Intelligence and Analysis for Business). He’s also the president of Advanced Analytics, which is a consultancy that he founded.

In this episode, Michael describes various risks and opportunities associated with the growing adoption of AI tools. In particular, he describes a scenario where AI contributes to income disparity at the societal level by replacing significantly more low-skilled jobs than it creates, which is described in the Nobel-prize-winning work of Daron Acemoglu at MIT. Michael also covers how existing marketing attribution methods can be blind to the upstream impact of Generative AI, and how SEO and PPC might be affected by LLM-based search.

Nachiket is the Global Head, Data & Analytics Engineering at Wayfair. In that role, he spearheaded the company’s global implementation of data mesh.

In this episode, Nachiket describes the four pillars of data mesh, and in doing so, he shares a rare front-line view of client-side operations involving data governance, including rich detail about how data mesh is implemented within an enterprise-class organization, what a successful implementation of that strategy requires, cultural implications, impact on data lineage, build vs buy considerations, and how that approach to data transforms everything from real-time decision making to Generative AI.

Zach is the Senior Director, Head of AI, Data Science and Analytics at Fortune Brands, a portfolio of brands that includes names like Yale, Master Lock, SentrySafe, Moen faucets, House of Rohl and Therma‑Tru, among others.

In this episode, Zach describes his journey from a PhD in Mathematics with an MBA, to the head of AI at Fortune Brands. One project he describes is a solution that detects residential water leaks and that automatically shuts off the water to protect the house. Another project involves sentiment analysis of customer reviews, implemented in a way that’s tailored to the specific needs of product managers. The conversation also covers enabling technologies, including multi-modal language models, causal inference, and digital twins. Zach also describes the road from feasibility testing to full deployment.

Pavel holds two gold medals, two silver medals and three bronze on Kaggle. He’s also the author of several well-known libraries, including one that has about 20,000 stars on GitHub. Currently, he's a member of the open source engineering team at Hugging Face.

This episode covers Pavel's journey from growing up in a small town in Russia, to becoming a Kaggle competition master and computer vision engineer at Hugging Face. The conversation also covers many insights about Hugging Face itself, including Hugging Face Hub, Enterprise Hub, the Transformers library, the Diffusers library for generating images, pilot projects (about 20-30 every six months), auto-training and showcasing models on Spaces, Remi Cadene's LeRobot, and Zero GPU. It also covers information about Hugging Face datasets, including a 15-trillion token dataset that's big enough to train a base NLP model.

This episode features a conversation with Max Mozgovoy, co-founder of VoxDiscover, which plans to use Gen AI to interview customers for user experience (UX) research done by app developers, or for customer lifecycle research, to build customer journey maps or personas, or to identify which features to add or change for the next version of product or service. 

The underlying idea of VoxDiscover is to use an AI voice agent to capture customer feedback right in the app or on a website, automating the insight as well, thereby achieving a much lower cost as compared to external service providers that have to pay customers to participate, while also transforming the discovery process into something that happens on an ongoing basis, rather than only during the app or product development phase, or at quarterly or annual or even longer intervals.

With a PhD in machine learning and a specialization in computer vision, Marwa has done deep learning projects across several industries and is currently at Alohi, a Switzerland-based startup with 4 million users of document solutions like Sign.Plus, Fax.Plus and Scan.Plus. In this episode, she describes her journey from formal studies to hands-on work in industry, to her present role as a primary engineer at Alohi. The conversation also touches on what it’s like to maintain a connection to a top university as a Research Fellow, while also working full-time at a startup. Marwa also discusses the challenge of staying current on new developments, because many papers have only marginal value, which makes it hard to quickly identify the important ones. The conversation concludes with a discussion about emerging technologies that Marwa considers to be quite promising in her field, especially diffusion models and LLM foundational vision models.

Iqbal holds a PhD in Computer Science and Engineering, with a specialization in computer algorithms and graph theory. He is currently the Research Data Science Director at the University of Arizona. Aside from his innovative achievement in creating a knowledge map for the university (kmap.arizona.edu), he has one of the most remarkable and inspiring stories you’ll ever hear about overcoming disadvantages. (He grew up in a rural part of Bangladesh and was in 11th grade before he ever saw electricity.)

In this episode, Iqbal describes the brief he was working on when he created the University of Arizona knowledge map. One of the interesting aspects of the project is that it is literally a map, with a visual “geography” of the various departments and disciplines within the university, laid out in a 2D plane, where the size and location of the polygons are based on the relative size and relationships between departments. Everything is fully interactive, including the ability to zoom in or zoom out, like one would do on a physical map, and where every element can be queried to extract more information from more than a dozen data sources, including both structured and unstructured data. The conversation also includes insights into how a large and complex project like this takes shape over time.

Jeffrey is one of those rare people who knew exactly what he wanted to do from an early age: Use mathematics to solve business problems. After earning his Masters Degree in Statistics from the University of Chicago, he went on to do high-impact analytics work in banking, retail and telco, holding senior leadership positions at top companies like Sally Beauty, HSN, AT&T and Bloomin' Brands. Currently, he leads Digital Analytics at JCPenney.

In this episode, Jeffrey shares a detailed and very insightful analysis of all the factors and trade-offs involved in making a build vs buy decision for analytics or AI solutions in business. He also names what he sees as the best applications for leveraging AI and machine learning in retail – including a few especially promising ones that he has his eye on right now. The show concludes with great career advice for any students who are currently enrolled in college-level data science or AI programs.

Tapan is the Executive Director of Marketing Sciences at OMD. This episode is a very thorough compendium of marketing science principles and practices, as applied to decisions about paid media. Topics include:

•    Marketing Mix Modeling (MMM) vs Multi-Touch Attribution (MTA) vs experiments, including how to use each and for what purpose •    KPIs: Incremental sales vs top-of-funnel metrics like awareness and consideration •    How to design incremental changes in the direction that a model is pointing •    Testing a new marketing mix in various geo markets •    Judicious use of model hyperparameters, such as AdStocks •    The best cadence for measurement •    Third-party cookies (Implications of the shift towards user choice) •    Recent tools: Google Meridian (a Bayesian approach) vs Meta Robyn (a frequentist approach)

These ideas are brought to life with an example from a financial services brand, planning a big launch, with a need to answer questions like: How should the paid media budget be allocated? Should the focus be branding first and performance later? How should ad spend be flighted (Always-On vs Burst and Trickle)?

In effect, this show is a Master Class regarding marketing science, as described by one of the world’s top authorities on the topic, who’s currently an Executive Director at OMD, which is the world’s largest media agency network.