As the Director of the AI Solutions Hub at the Institute for Experiential AI, I lead a team of data scientists and engineers who deliver Responsible AI solutions (both predictive and generative AI) at scale.

Machine Learning at Scale is an example course I developed and teach. This course teaches the underlying principles required to develop scalable machine learning pipelines for structured and unstructured data at the petabyte scale. Students will gain hands-on experience in Apache Hadoop and Apache Spark.https://www.ischool.berkeley.edu/courses/datasci/261https://www.ischool.berkeley.edu/people/james-shanahan

Architecting and engineering scalable and reliable systems requiring specialized knowledge of deep learning, computer vision, information retrieval (Web search), machine learning, data-mining, text-mining, recommender systems, and web metricsCurrent clients include Target (retail), ChartBoost (mobile digital advertising) Aylien (natural language processing as a service Collinear Networks (next generation, ultra-high capacity, over-the-air connectivity solutions for global communication networks) Experian (consumer credit reporting agency) Stealth mode startup in autonomous driving (object detection; driver monitoring)Previous clients include Nokia-Siemens Networks, Adobe, Akamai, AT&T Interactive, Digg.com, Ancestry.com, SearchMe.com, W3i.com, eBay, SkyGrid.com, Scoutlabs.com, MyOfferPal.com, Managing multiple teams of engineers and projects

Architected system for sponsoring articles embedded in news streamsDeveloped a news ranking system

• Consulting Executive Director of Research at AT&T Interactive; • Lead a team of 12 researchers (mainly PhDs)• Team focused on problems in query parsing, location disambiguation, learning to rank for local and mobile search, metrics and analytics• Published in top-tier conferences• Patented novel ideas in local search• Collaborated closely with AT&T Research Labs• Helped establish AT&T Research Labs (West Coast).

Architected the websearch platform at SearchMe Architected and developed the learning to rank framework at SearchMe Project managed engineering teams and editorial teams.

• Member of executive team where I collaborated with other execs to establish a product vision and road map to achieve the company’s goals and business strategies• Technical and scientific leader, focusing on the development and deployment of an online ad targeting system (CPA, CPC, CPM) in a principled and measured way that leverages advanced statistical and machine learning techniques• Lead strategic planning for technology and development initiatives • Evaluated and identified appropriate technology platforms• Provided leadership in defining product requirements and oversaw product development from design to execution• Attracted, coached, and motivated a world class team of software developers• Managed technical advisory board at Turn Inc. (that included Professor Jerry Friedman (Stanford University), Andreas Weigend and others)• Shared knowledge, mentored, and educated the organization’s investors, management, staff, partners, customers, and stakeholders with regard to the company’s technological vision, opportunities, and challenges• These responsibilities include leveraging the entire reservoir of data assets in order to develop methods for identifying key optimizations, deploying relevant analytical tools and improving the user experience

• Managed a team of research scientists and engineers• Lead strategic planning for technology and development initiatives • Attracted, coached, and motivated a world class team of software developers• Development of novel and proprietary online search and classification algorithms using support vector machines, probabilistic modeling, and committees of filters. • Initiated and developed a knowledge management system based on workflow mining from event streams (i.e., events can be both structured, e.g., stock price, and unstructured, e.g., free text such as a negative product review).• To date this work has resulted in the following publications: three books; two book chapters; eight peer-reviewed conference and workshop papers; four evaluation conference papers; three patent filings; and numerous technical reports in the field of information and knowledge management.

• Introduced and developed, in collaboration with Greg Grefenstette, a new information access paradigm centered on agentized documents: Document Souls. This work resulted in nine patent, and 5 million dollars of funding from Xerox Venture Fund. Document Souls is currently being commercialized at Xerox. • Co-PI of KerMIT project, a European Commission project in Kernel Methods for text and image classification, and clustering.• This work resulted in the following publications: a book; a book chapter; one evaluation conference paper; nine patent filings; and numerous technical reports in the field of information and knowledge management.

• Developed a new approach for object recognition in images, fuzzy perceptual organization. • This work resulted in the following publications: one journal paper; a book chapter; seven peer-reviewed conference and workshop papers in the field of image understanding.