We could not locate the page you were looking for.

Below we have generated a list of search results based on the page you were trying to reach.

404 Error
Friday’s TWIMLcon Executive Summit closed out a full first week at the conference! Speakers from BP, Walmart, Accenture, Qualcomm, Orangetheory Fitness, and more shared their experiences and insights on key issues faced by AI/ML leaders and teams. The day began with a keynote interview featuring Franziska Bell, VP of Data and Analytics at BP. Fran had some very strong advice on what it takes to ensure ML project success. Her principles include creating mutual partnership between the business and the data team early on in the process; working hard to ensure that the data team is actually solving the business need; and emphasizing the importance of empathy, understanding, and common goals and language among the cross-disciplinary teams building data products. The first panel of the day focused on Building the Business Case for ML Platforms and featured Divya Jain (Director of ML Platform, Adobe), Justin Norman (VP Data Science and Analytics at Yelp), and Kirk Borne (Principal Data Scientist and Executive Advisor, Booz Allen Hamilton). We discussed business value, measuring impact and ROI, build vs. buy, centralized vs. embedded teams, and standardization of infrastructure vs. flexibility. One attendee question prompted panelists to explore the topic of whether centralization was even a good thing. All panelists had strong opinions on this topic--not always in agreement--but Justin summarized it well with the following: “Businesses have many teams, those teams have requirements and those requirements should drive the platform choices. If it makes sense to centralize something... then do it. But if a team is doing something very unique with a different set of requirements than the other teams, they may need their own vertically integrated stack.” The next session had Adrian Cartier (VP of Data Science, Ocelot Consulting), Andy Minteer (Senior Director, Digital Transformation - Head AI Products, Walmart Global Tech), Jurgen Weichenberger (Data Science Senior Principal & Global AI Lead, Resources, Accenture) up to discuss Why ML projects Fail and How to Ensure Their Success. Right off the bat, Andrew challenged the idea of failure and had us rethink what failure even means. He asked: “What if the model is accurate but nobody adopts it? Isn’t that also failure?” Jurgen, who has worked with many customers in many industries, cautioned that it’s important to back up even further and to assess where the customer is on their maturity curve: Some industries are further ahead than others and that will drive a lot of what success and failure even mean to them. The panel closed with a discussion about the central role of people in the technology decisions leaders make. Jurgen offered: “It is our obligation to bring our customers on the journey with us. We need to be in the mindset that we are enabling people to do their jobs. You need to take the whole company on a journey with you... Bring them along, build trust and confidence, and show them how this can make their lives easier.” The fourth session of the day centered around what is required when Building Teams and Cultures that Support ML Innovation. For this discussion, we invited Ameen Kazerouni (Chief Analytics Officer, Orangetheory Fitness), Pardis Noorzad (Head of Data Science, Carbon Health), and Ziad Asghar (VP of AI at Qualcomm) to share their thoughts. The conversation included topics such as: what are the factors in building high-performance teams; how do we measure team success; and what is the role of culture in building teams. Sufficient budgets, common language, and shared rituals were all mentioned as key elements enabling effective teams. The impact of the pandemic on teams, namely the accelerated shift to remote work, was discussed as well. Ziad left us with this amusing thought on the topic: “If 2020 had a t-shirt, it would read: ‘Hey we can’t hear you, you’re on mute,’” illustrating how fundamental some of the challenges we face are. The final session of the day, the Executive Summit Roundtable Discussion, was a particularly animated and rich discussion. Ameen Kazerouni, Hussein Mehanna (VP, Head of ML/AI, Cruise), and Paul van der Boor (Senior Director of Data Science, Prosus Group), each shared their experiences on a topic relevant to leading ML teams and then Sam facilitated a great discussion afterwards with the attendees. A few of the many compelling ideas that came out of this section include: Ameen’s suggestion that: “The currency of an analytics team is trust, not data.” Paul’s insights from the experience of one of the teams at Prosus which has developed dashboards to granularly track the impact of every ML model they deliver on the business, and the need to understand whether a project’s key contribution is operational (improving what you already do) or innovation (doing new things). Hussein’s definition of “AI Native Products” as those that must leverage AI even at the MVP stage and his mind-blowing hypothesis (presented first at TWIMLcon!) that in order for organizations to create AI-Native Products that they need to organize internally like a neural network. We had a fun and engaging discussion after those three wrapped up and I think the summary was that we in the ML community have been spoiled with an explosion of new tools and techniques, and that at some point, there will likely be a “great reckoning” where the toolchain will all converge and MLOps will become more standardized. As one attendee, Gavin Bell, put it: “We used to have serial ports, parallel ports, printer ports, display ports, headphone jacks...and now it’s all USB-C. What is the respiration of this round of technology going to leave behind? “ Big thanks to Adrian, Ameen, Andy, Divya, Franziska, Jurgen, Justin, Kirk, Pardis, Paul, Ziad, Hussein and all of the Executive Summit attendees for a fun and stimulating day of discussion on these very important topics. And special thanks to Qualcomm, Executive Summit Platinum Sponsor. If you missed the session today, it’s not too late to register for TWIMLcon! There are still four more days of sessions next week. “Executive” tickets offer on-demand access to all of the Executive Summit sessions you missed, as well as the entirety of TWIMLcon. All tickets offer on-demand access to all regular conference sessions through the end of January, and “Pro Plus” and “Executive” tickets let you watch replay sessions whenever you like. You can check out the TWIMLcon agenda here and the speakers here. See you next week!
Today we are joined by Mark Ryan, author of Deep Learning with Structured Data. Mark started his journey as many thoughtful people do: trying to solve a problem. While working on the Support team at IBM Data and AI, he created a prototype deep learning model to predict the time of support ticket completion and which tickets would escalate. During this process, he saw that there was a lack of general structured data sets that people could apply their models to. As he was contemplating this problem, Mark noticed that in his hometown of Toronto, the streetcar network was causing problems through gridlock and delays. He created a deep learning model to predict these delays, but more importantly, gathered an open data set that was the perfect size and variety, and catapulted his contemplation into the book it is today. In this episode, Mark shares the benefits of applying deep learning to structured data (and recently reduced barriers to entry), details of his experience with a range of data sets, the everlasting appreciation he and Sam shares for the Fast.ai course by Jeremy Howard, and the contents of his new book, aimed to help set up and maintain deep learning models with structured data. If you would like early access to the book, please use the link below. You will have the opportunity to make comments, provide feedback and give recommendations that will be incorporated into the final published product. Check it out!
Last week we shared with excitement the news about a special Halloween event we were planning for October 30th in New York City. Well, due to unforeseen events beyond our control, the event is now cancelled. We were really looking forward to the event and are incredibly disappointed about its cancellation. If you purchased tickets via the eventbrite or splashthat pages you should have been automatically refunded.