Embedded Files
Adobe
Adobe
San Jose, CA, USA
San Jose, CA, USA
tracyhollowayking "at" gmail "at" com
tracyhollowayking "at" gmail "at" com
Engineering and applied science manager and individual contributor with an emphasis on the practical application of machine learning and natural language processing to large-scale production problems.
Adobe
Adobe
Senior Principal Scientist, Sensei and Search 12/2021-present
Senior Principal Scientist, Sensei and Search 12/2021-present
Senior principal scientist on the Adobe Sensei and Search team. Working across the Sensei and Search team to build state-of-the art search and natural language processing solutions for use across Adobe products.
Principal Scientist, Sensei and Search 02/2019-11/2021
Principal Scientist, Sensei and Search 02/2019-11/2021
Principal scientist on the Adobe Sensei and Search team. Working across the Sensei and Search team to build state-of-the art search and natural language processing solutions for use across Adobe products.
Amazon
Amazon
Senior Manager Sponsored Products Utilization, Amazon Sponsored Products 01/2018-02/2019
Senior Manager Sponsored Products Utilization, Amazon Sponsored Products 01/2018-02/2019
Applied science and engineering manager for the Utilization team in Amazon Sponsored Products (SP).
The SP Utilization team determines which sponsored product ads to show to a customer on Amazon search, product detail, and other on-Amazon pages. Sourcing determines the candidate set of ads, matching shopper intent (e.g. their search query) with advertiser intent (e.g. keywords they bid on). The auction determines which of the candidate ads to show using a query-time auction that takes into account the bid, relevance, and other factors.
Team: 40+ applied scientists, engineers, and product managers responsible for sourcing and auction. Worked closely with ad response prediction, economist, and infrastructure teams.
Defect reduction: Improved relevance of SP on search and product detail pages. Developed off-line and on-line ML/DL models to predict relevance, allowing for new, more aggressive sourcing strategies. Human judgment platform to track goals, make AB test decisions, and create training data.
Auction function: Collaborated with the economist team to incorporate publisher cost into the SP auction.
Bidding controls: Launched next generation of advertiser-facing bidding controls, including fixed and automated bidding strategies and placement-specific bids. Transferred post-launch to the SP advertiser controls team.
Technical Advisor Amazon Search and Visual Search 12/2017-01/2018
Technical Advisor Amazon Search and Visual Search 12/2017-01/2018
Senior Manager Query Intent and Matching, Amazon Search 12/2014-12/2017
Senior Manager Query Intent and Matching, Amazon Search 12/2014-12/2017
The Query Intent and Matching team determines the semantic intent of Amazon customers’ search queries and works across teams to provide search results and features reflecting these intents. Our goal is a full understanding of customer queries and a search experience that acts on this understanding with dialogue-based interaction when helpful to customers.
Team: Built team from 8 to 60 people (1/3 applied scientists, 2/3 SDEs) across three sites.
Query Understanding: Shallow and deep analysis of Amazon product search queries. Created QUZen service to serve Query Understanding throughout Amazon.
Defect Reduction: Used Query Understanding signals in matching and ranking to reduce Search defects (e.g. products that do not match the query intent) across marketplaces. Established metrics, weblab launch criteria, and measurement processes for search defects program.
Search suggestions: New auto-suggestion (inline search suggestions) service enabling weblabs for ranking and CX changes across all devices. Ranking and CX improvements. Created Guidance team to improve static and interactive navigation across Amazon Search.
Secondary Language Search (Mozart): Worked with Search, International and Browse to define customer experience, architect and launch search in secondary languages (e.g. Spanish in US).
eBay Inc.
eBay Inc.
Director of Engineering Search Science's refinements and universal search team: 2/2014-12/2014
Director of Engineering Search Science's refinements and universal search team: 2/2014-12/2014
The Refinements and Universal Search team guides customers through the shopping funnel to rapidly refine their queries and serendipitously discover new products. The team builds services and algorithms to power these experiences through front-end features.
Universal search platform: Design, metrics, and optimization algorithms for answers UX templates and placement within the search result set.
Structured data refinements: Algorithms and data to determine which structured data (categories and refinements) to display to customers.
RRelated search: Algorithm improvements including implementation of explore-exploit techniques and extending coverage to tail queries through soft match techniques.
IProduct captions: Item description summarization for search engine optimization (SEO) landing pages and product pages.
Principal Product Manager Search Science's query services and search spam team: 7/2011-2/2014
Principal Product Manager Search Science's query services and search spam team: 7/2011-2/2014
The Query Services and Spam team in the Search Science team is responsible for improving the quality of organic search results. The team accomplishes this through query understanding and query expansion, spelling correction, and the elimination of search spam including duplicate results, catalog spam, and search manipulation.
Query understanding and expansion: Algorithms and implementation to map from query words into synonyms and product-based structured data.
Fitment: Query triggering for motors parts and accessory search results and UX.
Spam classification: User-visible spam identification for use in ranking, recall, and customer service referrals. Improvements to the speed and reliability of search manipulation detection pipelines.
Search engine migration: Product strategy and execution for Search Science migration to new “Cassini” search engine.
Microsoft Corp.
Microsoft Corp.
Principal Development Manager BingSF's natural language engineering group: 6/2009-7/2011
Principal Development Manager BingSF's natural language engineering group: 6/2009-7/2011
The Natural Language Engineering team develops semantic natural language query and document processing for Bing search, focusing on improving search results for tail queries and answer triggering. Responsible for applied science and engineering groups focusing on statistical text processing, semantic analysis platform and models, tail query processing, summarization, and corpus development and evaluation.
Whole Page Relevance: Contributions to query processing team through query natural language processing including stemming, re-inflection, and core term identification.
Vertical experiences: For Bing’s Metro (hyperlocal) experience, delivered result ranking, query processing, data curation, SERP answers, and targeted metrics. For Commerce vertical, delivered components for review summarization, query classification and result ranking, and scalable answer triggering.
Natural language processing services: Ported Powerset shallow and deep natural language technology (tokenization, stemming, dependency structure, and semantics) to Windows and made these available as libraries and services to Bing feature teams.
Principal Development Lead BingSF's natural language engineering group: 10/2008-6/2009
Principal Development Lead BingSF's natural language engineering group: 10/2008-6/2009
The Natural Language Engineering team develops natural language query and document processing for the Wikipedia-based Reference vertical. Responsible for applied science and engineering teams.
Fact extraction: Increased precision of document and query processing and of tuple extraction for Factz feature. Established component and system level regression testing for accuracy and latency and conducted comparisons with competitive natural language technologies.
Palo Alto Research Center (PARC)
Palo Alto Research Center (PARC)
Area Manager and Senior Member of the research staff PARC's natural language theory and technology (NLTT) group: 7/2006-9/2008
Area Manager and Senior Member of the research staff PARC's natural language theory and technology (NLTT) group: 7/2006-9/2008
Member of the research staff PARC's natural language theory and technology (NLTT) group: 9/1997-6/2006; Consultant: 9/1995-8/1997
Member of the research staff PARC's natural language theory and technology (NLTT) group: 9/1997-6/2006; Consultant: 9/1995-8/1997
The Natural Language Theory and Technology team conducted research on deep natural language processing, i.e. mapping from text to semantic and knowledge representations. The team built large-scale prototype applications such as machine translation, question answering, and summarization to guide fundamental theoretical research. In addition to Xerox-internal projects, the team worked with academic institutions and industrial partners (e.g. the Parallel Grammar project).
Area manager responsibilities: Established research strategy including IP and field-of-use. Secured internal and external (government and industry) project funding. Balanced multiple project goals and deliverables. Received excellence in engineering and project performance awards from PARC.
Individual contributor responsibilities: Focus on grammar engineering (syntax and semantics) for multilingual applications. Developed broad-coverage deep English grammar and lexicon, including performance tuning and evaluation.
Patents
Patents
Query expansion classifier for e-commerce
Patent number: 9135330; Issued: 2015; Assignee: eBay, Inc.
Inventors: Ravi Chandra Jammalamadaka, Vamsi Krishna Salaka, Brian Scott Johnson, Tracy Holloway King
Generating snippets based on content features
Patent number: 8788260; Issued: 2014; Assignee: Microsoft Corporation
Inventors: Valerie Rose Nygaard, Riccardo Turchetto, Joanna Mun Yee Chan, Christian Biemann, David Dongjah Ahn, Andrea Ryerson Burbank, Feng Pan, Timothy McDonnell Converse, James Michael Reinhold, Tracy Holloway King
Using a content database to infer context information for activities from messages
Patent number: 8661046; Issued: 2014; Assignee: Palo Alto Research Center Incorporated
Inventors: Tracy Holloway King, Kurt E. Partridge, Nicolas Ducheneaut, Ji Fang
Non-sensitive-passage database for cut-and-paste attack detection systems
Patent number: 8402542; Issued: 2013; Assignee: Palo Alto Research Center Incorporated
Inventors: Tracy H. King, Philippe J. P. Golle, John T. Maxwell, III, Jessica N. Staddon
Method and apparatus for detecting sensitive content in a document
Patent number: 8271483; Issued: 2012; Assignee: Palo Alto Research Center Incorporated
Inventors: Jessica N. Staddon, Richard Chow, Valeria de Paiva, Philippe J. P. Golle, Ji Fang, Tracy Holloway King
Systems and methods for detecting entailment and contradiction
Patent number: 7313515; Issued: 2007; Assignee: Palo Alto Research Center Incorporated
Inventors: Richard S. Crouch, Tracy Holloway King
Education
Education
Stanford University: PhD 1993, Linguistics
Massachusetts Institute of Technology: BS 1988
Google Sites
Report abuse