Developing ML-driven seller personalization and commerce intelligence features that serve Square's overarching purpose of economic empowerment.

Director of an annual seminar on technology and society:• 2017: "New technologies to help blind people" • 2015: "New technologies and autism spectrum disorders"• 2014: "The transparent brain" with K. Zilles (Aachen), A. Sandberg (Oxford.), K. Szymielewicz (Panoptykon), G. Chalkiadakis (Crete)• 2013: "Visualizing the world" with Stephen Macknick (Scientific American), Susana Martinez Conde (Barrow Neurological Institute), Marco Bagni (Ixtract), Anna Pujadas (UPF)• 2012: "Memory and the formation of individual and collective identities" with Alison Winter (Chicago), Martin Conway (City London), Josep Fontana (UPF), Albert Lladó (Hospital Clínic)• 2011: "Why music? An interdisciplinary conversation" with Philip Ball, Stefan Koelsch, Kevin Kleinmann, Yannie Tan, Adolf Pla, Hèctor Parra, Joan Vives• 2010: "The impact of the digital revolution on the human condition" with Gary Wolf, Ch. Pentzold, F.S. Caballero• 2009: "(Ir)rational choices: The new models of human behavior, economy, and happiness" with B. Frey, T. Gilovich, E. Shafir, A. Clark• 2008: "The nature of time, from physics to philosophy" with Huw Price, David Eagleman, D. Bueti, Stefan Klein• 2007 "Neuroethics: Law and the neuroscience of free will" with Laurence Tancredi, Patricia S. Churchland, Thomas Metzinger, Daniel Wegner• 2006: "Machine learning for language technologies" with R. Mooney, F. Casacuberta, W. Daelemans, L. Wanner, S. Clark• 2005: "Societal applications and repercussions of language technologies" with Orkut Buyukkokten, J. Abaitua, M. Hoffman, A. Virtanen• 2004: "Recent advances in language technologies" with Richard Stern, Bob Carpenter, Harold Somers• 2003: "Information retrieval" with Ed Hovy, Vibhu Mittal, Mehran Sahami, F. Sebastiani, L. Màrquez• 2002: "Machine translation" with J.Carbonell, A.Trujillo, S.Nirenburg, J. Hutchins• 2001: "Speech recognition" with C.Nadeu, M.Woszczyna, XD.Huang• 2000: "Language technologies" with A.Black, A.Trujillo, A.Waibel

Drank from the Yak firehose, i.e., developed machine intelligence services that analyze text, images, and user actions at scale to classify content, discover topics and trends, and make personalized recommendations.

Conducted research in speech recognition, language understanding, information retrieval, and machine learning applied to the MindMeld anticipatory computing platform. Acquired by Cisco for $125M.

Led Verint's global speech and text analytics R&D.

Led Nexidia's core technology team. Conducted original research in audio and speech technologies that power disruptive applications at scale in call center, intelligence, and media markets. Executed on the mission to take innovative ideas from inside and outside the company, evaluate their technical and business potential, rapidly build prototypes, and turn the most promising ones into new features, products, and even lines of business.

Led all stages of planning, design, and implementation of Freedom Navigator, a robust and scalable speech analytics solution that mines large volumes of audio to extract actionable business intelligence, acquiring ca. 200 customers and achieving over $1M in revenues in less than two years.Freedom Navigator achieves scalability by distributing computationally-intensive tasks over a grid of blade servers, configuration-free setup by automatic, dynamic discovery of available servers, high throughput of keyword searches by a bookkeeping mechanism that maximizes the use of indexed files cached locally on each blade, and flexibility by providing an XML-based API (NavigatorXML) and a web-based, thin-client interface for end users.Also very much involved in marketing and sales, including visits to Fortune 100 companies to present and demonstrate the product suite.Filed patent as sole inventor of "A system and method for improving the accuracy of audio searching".

Designed and developed SOUP, a stochastic, chart-based, top-down parser, especially engineered for real-time analysis of spoken language with very large, multi-domain semantic grammars. SOUP achieves flexibility by encoding context-free grammars written in a standard formalism, such as JSGF, as probabilistic recursive transition networks that can be dynamically modified, and robustness by allowing the skipping of input words at any position and producing ranked interpretations that may consist of multiple parse trees.

Developed the language-identification algorithm that shipped in Microsoft Word 2000.