Hi Everyone,
As a follow up to our October 25, 2018 meeting, we would like to extend a tremendous thank you to our esteemed speakers, Mark Donner, N.Y. Site Director and Head of SRE at Uber, Stephen Brobst, Truision Partner, CTO of Teradata Corp., & Managing Partner at Sampo Technologies & Systems, and Thomas Fay, Senior Vice President of Enterprise Architecture at Nasdaq. Special thanks to Mark Donner of Uber for hosting our meeting, and providing a wonderful venue for our event.
The following companies are currently registered for the October 2018 meeting:
Uber, Nasdaq, Moore Capital Management, Voya Financial, Promontory Financial Group, Broadridge FMarc Donner was born in a log cabin inancial Solutions, Discovery Capital Management, LLC, Kita Capital Management, LLC, SIM, Aflac, SpringHarbor Financial Group LLC, AEGIS, CDPHP, Fred Alger Management Inc., High 5 Games, Nine West Holdings, The Andrew W. Mellon Foundation, Wiley & Sons, iQ Venture Advisors, L.P., St. John’s University, UTIMCO, Contrarian Capital, Imagineer Technology Group, Gumtree Tech, The Carlyle Group, Swift, L’Oreal, Global Tech Solutions, Vikar Technologies, New York Blood Center, Gerber Life Insurance Company, The Leading Hotels of the World, Principis Capital, Elliott Management Corporation, Centurion Global Management, LLC, The Joint Industry Board, Financial Guaranty Insurance Company, The Financial Risk Group, JP Morgan Chase, Jensen Precast and Luminus Management LLC.
Meeting Summary:
Mark Donner, N.Y. Site Director and Head of SRE at Uber
The meeting was opened by Mark Donner who shared a brief overview of the NY Uber operations and business activities:
- UberEATS – this segment of their business is growing faster than the ride share business grew. It is a three-way match between the courier, the restaurant/provider and the customer
- Jump Bikes & Lime Scooters – Uber bought this business which is being run from the NY office
- Uber Freight – also growing faster than the original ride share segment
- Uber Health – a new segment that will replace ambulance transportation to the hospital or other medical appointments, patient drop-off and pick-up by specially trained drivers & data protection to conform with HIPAA regulations
- Uber experience is also work that is done in the NY Offices
- U4B – Uber for Business sales team in in NY office
Stephen Brobst, Truision Partner, CTO of Teradata Corp., & Managing Partner at Sampo Technologies & Systems, discussed Artificial Intelligence.
Stephen Brobst spoke on Machine Learning and Artificial Intelligence, here are the highlights from his talk:
- AI is over-hyped right now with extremely big expectations. AI is off the charts, much as Hadoop was five years ago. No solution solves all problems. Sales people, IT specialists, vendors are lying and sending companies into a trough of disillusionment.
- There are very large data sets – hairballs of data that currently require resources spending 90% of their time sifting through the data vs. 10% of their time analyzing the data.
- The current data mining process is a very large, long process that results in the algorithms and predictive models being out-of-date shortly after the model is complete. Measuring the model is very manual and happens infrequently; (once every 6 months). This creates a saw-tooth process.
- A better approach is to constantly tweak the algorithm that keeps the algorithm at the top of its performance.
- Setting realistic expectations is key. Gartner reports that by 2020 30% of CIO’s say that AI will be in the top 5 priorities of big enterprises. 75% of CIO’s have AI spending as one of their main goals. CEO’s are promising AI will be the answer to everything. Consolidation, however, is far off. If one is an investor, plan for industry shifting still ahead.
- Three domains have high value practical applications for AI deep learning: Predicting Demand (recommendations), Fraud Detection, and Failure Detection.
- Why is the hype happening now? New algorithms were introduced and GPU capabilities can be used for processing massive data programming code.
- Deep leaning is not the answer to everything. Choose wisely between deep learning and shallow (wide) learning which is more cost effective in many cases.
- The Industry & new players will evolve over the next 5-7years. There are still difficulties in utilizing these types of “black box” algorithms, for example, in fraud and risk management, showing how and why a bank will take action must be easy to explain to regulators.
Recommendations: Deep learning is for projects that are hungry for data, however It will take five to eight years before it stabilizes. Expect three more significant changes, and remember it is not the solution for all problems. Lastly, combine shallow and deep learning to maximize the technology for the best outcome.
Thomas Fay, Senior Vice President of Enterprise Architecture at Nasdaq, provided us with an interesting look at Quantum Computing.
Tom Fay from NASDAQ was the final presenter, who spoke about Quantum Computing. Here are some highlights from his talk:
- Quantum computing is accelerating, with practical uses emerging. It will be readily available as a cloud service in about 5 years. There is a race to supremacy happening amongst the key players in the field.
- Quantum Computing does not replace classic computing. It does not solve all problems, and is not programmed the same way as classical computing. It is not science fiction although challenges remain. There are significant advances being made in engineering labs around the globe.
- Quantum Computing does solve interesting problems. It is faster for solving some problems, and requires a unique skill set.
- How do Classic Computing compare to Quantum Computing:
| CLASSIC | QUANTUM |
| Classic physics | Quantum physics |
| O and 1 bits | Qubits |
| Sequential processing | Exponential processing |
| Deterministic | Probabilistic |
| Measurement is instructive | Measurement is destructive |
- There are three basic concepts – superposition, entanglement, collapse.
- Classic problems Quantum Computing can solve – cryptography, model optimization, machine learning and prediction acceleration (unsupervised learning), and searching for big data.
- Why do we care in financial services? Quantum Computing is critical in data privacy, simulating market conditions, finding alpha (predicting effect on market structure), and credit scoring.
- Money is pouring in to the field with traction growing. Key players in 2017 are Google, IBM, Microsoft, Intel and Rigetti.
- Recent headlines on Quantum Computing highlight a talent shortage, as the skill set is hard to find. Companies are beginning to offer cloud platforms for running Quantum Computing – Quantum Computing as a Service.
- Some challenges remain: Processing “burns” qubits for error correction, and error rates have to come down from the current 100 to 1. Microsoft is working on a technology that they hope to achieve an order of magnitude improvement.
Summary: In the near future there will be problems that even faulty Quantum Computing can solve. The Quantum Computing race to supremacy has led to a boom in quasi-quantum-classical computing algorithms. The Hybrid approach (quantum & classical) allows for algorithms that leverage quantum computing to achieve a quantum advantage.
Final Observations:
Our fabulous speakers provided us with interesting insight as to what our future will bring. As the questions continued late into the evening, it was clear that many more remain to be answered. Thanks again to Stephen and Tom, and to Marc Donner, our host at Uber for helping us create a truly exceptional event! To all those of you who joined us, thank you for your interest and incredible participation! This was truly an amazing evening!
Wishing you all a wonderful holiday season!
-malka
Malka Treuhaft
Executive Director East Coast CIO Forum &
President
Truision Inc.
646.942.2625 (office)
917.589.1069 (mobile)
718.375.1529 (fax)
Stephen Brobst’s Bio
Stephen Brobst is a Managing Partner at Sampo Technologies & Systems and also serves as the Chief Technology Officer for Teradata Corporation. Stephen performed his graduate work in Computer Science at the Massachusetts Institute of Technology where his Masters and PhD research focused on high-performance parallel processing. He also completed an MBA with joint course and thesis work at the Harvard Business School and the MIT Sloan School of Management. Stephen is a TDWI Fellow and has been on the faculty of The Data Warehousing Institute since 1996. During Barack Obama’s first term he was also appointed to the Presidential Council of Advisors on Science and Technology (PCAST) in the working group on Networking and Information Technology Research and Development (NITRD) where he worked on development of the Big Data strategy for the US government. In 2014 he was ranked by ExecRank as the #4 CTO in the United States (behind the CTOs from Amazon.com, Tesla Motors, and Intel) out of a pool of 10,000+ CTOs.
Thomas Fay’s Bio
Tom Fay is Senior Vice President of Enterprise Architecture at Nasdaq. In this multi-disciplinary role, Mr. Fay is responsible for the architecture, design, and performance of critical technology solutions across the company. Since rejoining Nasdaq in 2010, Mr. Fay has also led the System Performance and Engineering Group, which was responsible for increasing performance of the US based exchanges. Under his leadership, the responsibilities of the group expanded to include FPGA development, Capacity Management, Platform development and R&D, leading to the formal creation of the Enterprise Architecture Group in 2014. Mr. Fay’s experience spans over thirty years and he has held senior technology positions in a variety of industries including telecommunications, defense and financial services. From 2008 to 2010, Mr. Fay was Chief Technology Officer for Virtu Financial and was responsible for the Firm’s trading system including network infrastructure, proprietary software development and back office database systems. Mr. Fay originally joined Nasdaq as part of the company’s acquisition of INET in 2005. During this time, he was responsible for the consolidation of multiple trading platforms into the current INET technology based single book system, the development of Nasdaq’s US Equities Clearing System, and creation of an automated regression test framework to verify functional changes to the trading systems. Mr. Fay is a graduate of Stevens Institute of Technology.
Marc Donner’s Bio
Marc Donner was born in a log cabin on the Lower East Side of Manhattan in the second half of the twentieth century. After high school he escaped to Los Angeles where he studied physics, math, and electrical engineering at Caltech. Later he worked at NASA on planetary radar. He got lost one day and ended up back on the east coast where he joined IBM Research for a while before finding his way to CMU where he earned a PhD in computer science and party organizing while programming Ivan Sutherland’s six-legged walking robot. A talk on juggling by Claude Shannon inspired him to build a juggling robot. While that work was going on he sponsored Ted Selker’s work that led to the TrackPoint. He then got involved in large-scale distributed computing which led him from IBM to Morgan Stanley. There his possession of the root password let him build the first Intranet. Later on he fooled around with big data in the marketing department and built some entertaining economic and financial models in the research department. After his spider sense warned him of the impending collapse of the financial industry, he moved to Google where he led a number of efforts in the New York office – Ads development, DoubleClick integration, Google Health, Google Finance, Network Software, and the Google Art Project. After his spider sense tingled again, he moved to Uber as the Engineering Site Director for the New York office, after a two-year relapse with the financial services industry. The simple life lasted only six months, after which Uber asked him to take on leadership of all of SRE, which he then undertook
