Industry Roundtable: AI Data Curation
The adoption of Generative AI is moving IT operations teams into a new realm that involves processing a massive volume of many varieties of data, which imposes new demands on data scientists, data engineers and AI application developers. AIOps has been driven primarily by time-series observability data sourced from network, security and application infrastructure, and while data preparation and “wrangling” has always been a challenge, GenAI raises the bar to new level. Participate in an open discussion of the challenges and opportunities related to AI data curation in this new era. Is data curation for GenAI fundamentally different than for traditional AIOps? What are the key concerns and risks? How is data curation integral to AI model training and continuous tuning? Are there parallels with the DevOps CI/CD methodology?
This roundtable session, open to both IT professionals and suppliers, is subject to Chatham House rules.
Phil Tee is a VP Strategy and Operations, ISG AIOps at Dell and the former CEO and co-founder of Moogsoft, the pioneer and leading provider of artificial intelligence for IT operations (AIOps). Phil is a serial entrepreneur, inventor, and expert in IT service assurance — a category he was instrumental in shaping through his invention of Netcool (acquired by IBM) nearly 25 years ago. Within 5 years of graduating from Sussex, he co-founded Omnibus Transport Technologies Limited (OTT) to build and market Netcool/Omnibus. Since then, Phil has led numerous companies to successful exits, including RiverSoft (IPO) and Njini (acquired by Riverbed).
More recently at Moogsoft that he co-founded in 2011, he invented the technology that supports over 200 customers worldwide, including SAP, American Airlines, Yahoo!, Verizon, Outsystems, BNYM and HCL Technologies. Moogsoft has 72 patents and more than two dozen peer reviewed publications covering the technology, many of them authored or co-authored by Phil. Phil holds a PhD in Informatics and is also a Royal Society Entrepreneur in Residence.
Helen Gu is a leading expert in the field of distributed systems and predictive analytics. She is also Professor at North Carolina State University and leads a research group supported by over $4.2 million research grants from NSF, NSA, ARO, Google, IBM and Credit Suisse. She published more than 80 research papers and has filed 11 patents (10 of them have been issued).
Helen was on sabbatical at Google as a Visiting Scientist in 2015. She also worked at IBM T. J. Watson research center as a Research Staff Member in 2004-2007 working on the IBM stream processing system.
Helen Gu received her PhD in CS from UIUC in 2004 and BS in CS from Peking University in 1999. Her work has been widely reported by presses including NSF research highlights, Communications of ACM, and The Register, and won several best paper awards from prestigious International conferences including 10-year best paper award from prestigious ACM Symposium on Cloud Computing on her innovative work on predictive auto-scaling in cloud.
View Research Papers
Julian Guthrie is the CEO and founder of Alphy and co-inventor of Reflect AI, a proprietary technology that detects and flags harmful, unlawful, and unethical language in digital
communication in real time. Reflect AI integrates with everyday workplace tools to reduce the risk of costly litigation and reputational damage stemming from harmful communication.
Before founding Alphy in 2020, Julian spent twenty years as a journalist at the San Francisco Chronicle. Her journalism was nominated multiple times for the Pulitzer Prize. She has written
five nonfiction books, three of which became New York Times bestsellers. Her fourth book, Alpha Girls – being adapted for a television series – features four women who succeeded in the
high-stakes, heavily male world of venture capital in Silicon Valley and helped build some of the foremost companies of our time. Julian and her team have applied their language skills to build a
highly accurate and specialized AI communication classifier that detects over 30 categories of harm with a 92.4% accuracy detection rate.
