Resume

A dazzling and beautiful professional headshot Photo of Patrick Wagstrom

Patrick Wagstrom

Data Science and Artificial Intelligence Leader

Executive leader of cutting edge data science, machine learning, and software engineering organizations. I accelerate value from data by combining best practices from machine learning, risk management, cloud deployment, devops, and human-computer interaction. Hands-on technical leader that builds cohesive distributed teams by bringing together great people from diverse backgrounds while creating a shared sense of vision and ownership.

Professional Experience

    • Brightcove Boston, MA and Hartford, CT Chief Data Officer May 2021 - December 2022
      • Built and managed an organizaiton of 34 FTE and 12 contractors in engineering, data science, and product management responsible for data platforms, analytics and insights, data and model governance, and customer facing data products.
      • Executive sponsor and owner of the $12.3mm acquisition of Wicket Labs - Brightcove's biggest acquisition in 8 years that brought critical media analytics skills and focus to Brightcove. Brought the product to more the 50x as many customers in under 1 year.
      • Ownership multi-million dollar annual relationship with Google for Google Cloud Platform.
      • Led team and built out models of video classification from text transcripts, audio, and video processing - including sourcing of training data, development of algorithms, and creation of model execution platform.
      • Responsible for end-to-end analytics collection and processing system for tens of thousands of events a second, hundreds of millions of videos, billions of monthly views, and trillions of rows of data in our data warehouse.
      • Rearchitected and led migration of Brightcove's real-time analytics pipeline from Google Dataflow to Google Pub/Sub, resulting in a savings of $300k annually.
    • Verizon Basking Ridge, NJ and Hartford, CT Director of Emerging Technology September 2019-April 2021
      • Built a team 15 FTEs and additional 51 vendor contractors with a $15mm annual budget from the ground up - including attracting internal and external talent and negotiating for budget allocations.
      • Defined and architect overall enterprise strategy for reproducible machine learning and MLOps including data management, model training, model serving, model refit, and model risk management.
      • Led the advanced data warehouse and data pipeline architecture functions - migrating from a number of on prem ad-hoc solutions (i.e. Teradata and Hadoop) to a streaming cloud based solution around Google BigQuery.
      • Led proof-of-concept solutions around blockchain identity and asset management, augmented and virtual reality for Verizon retail and 5G UWB Home Internet, cross-device biometrics, and more.
      • Conceived of and built models for cell site energy efficiency resulting in a $5mm annual savings and conversion into a digital twin for future simulation.
      • Led team that developed a holistic model for connecting online brand discussion to customer profiles to support cusomter servicing - including customer and agent action prediction models and patent pending bot detection model.
      • Created and ran numerous cultural and knowledge sharing events including Data Science Reading Group and Emerging Tech Brown Bag Sessions along with heavy participation in internal and external speaking and outreach events.
      • Worked with legal and policy experts for the creation of numerous Verizon policies, including facial racognition, energy efficiency, and data privacy.
    • Capital One New York, NY and McLean, VA Senior Director of Data Science for Machine Learning Platforms January 2019-September 2019 Director of Data Science for Machine Intelligence November 2016-January 2019
      • Architect and overall lead for the Capital One Card Machine Learning Platform - a platform for scalable and personalized real-time reinforcement machine learning models on top of Kubernetes, gRPC, and AWS inside a regulated industry.
      • Led a distributed team of data scientists (12), data engineers (22), and product managers (3) to deliver our platform in a complex cloud-based enterprise environment.
      • Led a distributed team to build robust machine learning models that incorporate both unstructured text and structured analytical data to identify customer issues in a large corpus of customer communication.
      • Managed, grew, and nurtured a team of 18 data scientists across four states.
      • Built out an academic research partnership with universities to explore implications of fairness in artificial intelligence.
      • Conducted more than 200 interviews of data scientists, data engineers, and product managers to help grow Card Machine Learning from 33 to 168 people and Capital One's New York Card team from 8 to nearly 200.
    • IBM Watson Littleton, MA Research Staff Member/Technical Lead January 2015-November 2016
      • Led a globally distributed team that built the IBM Watson Conversation service - an innovative service to create rich interactions using natural language processing, entity recognition, and scripted dialog.
      • Global team lead for Watson Developer Cloud Tooling. Created cutting edge applications for creating, training, and maintaining cognitive and machine learning solutions including Watson Engagement Advisor and IBM Watson Natural Language Classifier.
      • Conducted research on optimal methods to build and configure cognitive and machine learning systems resulting in a 95% reduction in human time required to train these systems.
      • Presented cognitive solutions and technologies to more than a dozen companies in six countries.
      • Developed best-of-breed modern web application architecture based on Angular, Node.js, GitHub Enterprise, Slack, UrbanCode Deploy and more.
    • IBM Watson - Watson Life Yorktown Heights, NY Research Staff Member January 2014-January 2015
      • Leadership team member responsible for evaluating promising consumer applications of cognitive computing and guiding teams in lean startup and IBM design thinking processes to explore ideas.
      • Organized and led globally distributed teams on how to contextualize solutions to cognitive computing challenges.
      • Project lead for the IBM Food Truck at SXSW which demonstrated cognitive computing to more than 4,000 people and resulted in more than 1 billion media impressions.
      • Planned and executed international workshops on design thinking methodologies.
      • Implemented best practices for DevOps using IBM DevOps services, IBM BlueMix, Jenkins, and more.
    • IBM TJ Watson Research Center Yorktown Heights, NY Research Staff Member August 2009-January 2014
      • Analytics lead for JazzHub, IBM's cloud software development strategy. Designed analytics strategy, implemented metrics driven development, introduced A/B testing, and developed analytics dashboards.
      • Technical and strategy contributor to IBM's research and acquisition strategy for software development tools and processes.
      • Developed and designed GitMiner - an open source project used by 15 universities to perform graph analysis on large scale software engineering databases such as GitHub and BitBucket.
      • Led a research team to evaluate productivity of new users and small teams using IBM's enterprise software engineering and product development environments.
      • Published papers on topics around distributed collaboration, technical debt in software, and flow of ideas in software engineering communities.
      • Developed WhatsMyBrand, a framework for assessing an individual's personal brand by analyzing connections and contents of their actions through public social networks and relating those actions to the actions of others in their network.
      • Worked with IBM clients to teach about uncertainty and value elicitation in software development
      • Mapped extended stakeholders in enterprise software development and analyzed their relation to technical debt
      • Developed novel methods and metrics for understanding extended enterprise software development stakeholder collaboration and coordination
      • Managed research on interactions around software development with three different universities through an Open Collaborative Research grant
      • Mentored three Ph.D. level student interns on projects related to collaboration in software engineering
    • Carnegie Mellon University Pittsburgh, PA Graduate Research Assistant August 2003-July 2009
      • Designed and developed CVSMiner an open source tool to perform social network and technical analysis of software engineering ecosystems such as GNOME and Eclipse.
      • Worked with members of the GNOME Foundation and Eclipse Foundation to evaluate and improve the relationships between non-profit foundations that manage open source ecosystems and commercial firms.
      • Delivered lectures in classes on software engineering and technology policy.
      • Utilized a variety of qualitative and quantitative research methods: stakeholder interviews, message analysis, natural language processing, data mining, machine learning, and social network analysis to generate insight into largely ad hoc software development processes.
    • IBM TJ Watson Research Center Hawthorne, NY Summer Research Intern June 2007-August 2007
      • Expanded the Socio-Technical Congruence metric, which relates communication between individuals and technical dependencies inferred from archived data.
      • Developed a model of successful projects that transitioned from IBM proprietary technologies to strong open source communities.
      • Developed novel visualizations of communication and congruence in software development teams.
    • Computer Science Department, Illinois Institute of Technology Chicago, IL Teaching/Research Assistant September 2000-August 2003
      • Managed a group of twelve undergraduates to develop an ambitious automated tour system utilizing Segways and mobile devices for wireless location sensing and data in an era years before the iPhone and Android.
      • Hired and managed three undergraduates to develop a python based framework for pervasive computing based on web services technologies.
      • Designed graduate level class on grid and pervasive computing.
      • Delivered numerous lectures on a wide variety of related to distributed computer, operating systems, and computer architecture.
    • Argonne National Laboratory - Math and Computer Science Division Summer Research Intern Argonne, IL April 2002-September 2002
      • Developed the Grid Services Flow Language for specifying dependencies and flow in the GLOBUS environment.
      • Worked with the SciDAC Java CoG Kit Team and the Collaboratory for Multiscale Chemistry to develop a grid services based system for analysis of thermochemical tables.
    • LEC, Ltd Chicago, IL Senior Developer April 1999-September 2000
      • Designed, ordered, installed, and managed a commercial grade data center for advertising agency clients.
      • Architected and developed E-Stakes, a multi-million user capable system for tying offline purchases to online activities.
      • Designed and managed the technical components of the Chicago Transit Authority's "Take it and Win" promotion that utilized CTA transit cards to tie together offline and online behavior of transit riders.
      • Worked directly with designers and clients to sell and develop usable, novel, and cutting edge web experiences.
    • MyPoints Schaumburg, IL Developer April 1998-September 1998
      • Designed and implemented a complete customer relationship management system in PL/SQL and Java.
      • Integrated customer service system to work with multiple advertising campaigns and custom co-branded sites.

Education