Vector Institute Establishes New AI Engineering Team to Ramp up Applied AI for Sponsors and Partners

August 26, 2020

Vector Institute Establishes New AI Engineering Team to Ramp up Applied AI for Sponsors and Partners

August 26, 2020

The Vector Institute is establishing an AI engineering team, comprising highly-skilled technical staff focused on transforming cutting-edge research into reference applications, tools, frameworks, and model templates that will serve as foundations for new industrial applications and health care initiatives. The team will also provide industry and health care partners with the know-how and hands-on experience to adapt and operationalize these models within their organizations. New Vector executive Ron Bodkin will lead the team. Ron most recently held the role of Technical Director for Applied Artificial Intelligence in Google’s Cloud CTO office in Silicon Valley, where he focused on responsible AI and industry AI solutions, working with Google product teams and Google Cloud’s clients. He will assume the role of VP of AI Engineering and CIO at Vector and Engineering Lead at the Schwartz Reisman Institute for Technology and Society.

The practice of applying AI is a major priority in Vector’s new three-year strategy. The growth of Vector’s unique community of top researchers, leading companies, and health care organizations has enabled us to take this novel approach among research institutes and bridge the gap between fast-moving AI research and the needs of organizations that can benefit from putting applied AI into production.

Vector’s new AI engineering function
The AI engineering team will transform research into foundations for practical application and will continue to build out Vector’s compute resources and software stack to support research and industry projects. The team will focus on five domains:

● Industry
● Responsible AI
● Health
● Research
● AI infrastructure

Vector’s AI engineering team will deliver state-of-the-art tools and perform technical facilitation to support industry sponsors participating in existing knowledge transfer programs and 10 applied AI projects, as described in Vector’s three-year strategy. These applied projects ― in the mold of our recent NLP Project (aka Recreation of Large Scale Pre-trained Transformer Based Models) and newly-launched Dataset Shift project ― will enable industry sponsors to train cutting-edge AI models in Vector’s compute environment with the aim of gaining the hands-on experience necessary to operationalize them within their own organizations. Vector’s engineering team will contribute reusable software and mentoring to facilitate applying AI research into operational software. The team will also contribute to our Face-to-Face program, one-on-one meetings with researchers and engineering staff in which sponsors receive advice and feedback on AI problems, products, or processes specific to their respective organizations.

Responsible AI
Vector’s AI engineering team, in collaboration with the Schwartz Reisman Institute for Technology and Society, will focus on developing technology and frameworks that build responsible AI into model operationalization and address common challenges Vector partners face, including those related to the technical elements of AI fairness, model transparency, and governance. This will support sponsors to develop effective practices that enhance trust and lead to beneficial outcomes from AI systems.

Vector’s AI engineering team will deliver innovative AI solutions for health care and support health organizations to operationalize AI with a strong focus on privacy, technology modernization, and collaboration with practitioners. This will complement and build on Vector’s commitment to modernizing health data governance for research by enhancing the impact of our recent contribution of compute infrastructure to support the Ontario Health Data Platform for COVID-19 research, our collaborations with Unity Health Toronto and TAHSN hospitals to derive COVID-19 insights from hospital data, and our Health AI Data Access Platform (HAIDAP) collaboration with the Institute for Clinical Evaluative Sciences (ICES), UHN, and HPC4Health at The Hospital for Sick Children.

The AI engineering team will enhance and accelerate our researchers’ ability to conduct and reproduce experiments and push forward the frontier of AI research. AI engineering will make computations more efficient, more effectively parallel, and more easily deconstructed for performance debugging. With strong AI engineering tools, AI modellers can spend more time on innovative models and optimizations without risking ineffective use of resources or unrepeatable results. AI engineering will also foster the most promising innovations and seek to demonstrate their implications for AI practice, including through open source publication and maintenance.

AI Infrastructure
Vector’s AI engineering team will procure and operate one of the world’s most significant non-profit machine learning computational infrastructures, which will include thousands of GPUs, petaflops of computation, large scale data management, modern training frameworks, experiment workflow tools, and prioritized fairshare scheduling. The infrastructure will be quickly updated as computational technology innovations and open source community software developments occur.

Meet Ron Bodkin: Serial entrepreneur and seasoned technology executive focusing on responsible AI

After spending the majority of his professional life in the United States, Ron is repatriating to Canada to take this role at the Vector Institute. In his most recent role as Technical Director on the Applied Artificial Intelligence team in Google’s Cloud CTO office, Ron led strategic initiatives working with Google Cloud’s customers and partners, and worked with product teams on explainable AI and fairness capabilities in Google Cloud products.

Ron is also a serial entrepreneur. He was the founding CEO of Think Big Analytics, a company which developed corporate data strategies for clients, and provided integrated machine learning and data engineering services. In 2014, Think Big Analytics was acquired by Teradata, where Ron led Think Big as a business unit focused on analytic solutions, deep learning, and GPU compute. He later became Teradata’s Vice President & General Manager of Artificial Intelligence, leading a team with a charter to identify and commercialize machine learning products.

Ron’s mix of entrepreneurship, product development, machine learning, and responsible AI experience augments the Vector Institute’s credentials and capabilities as a world-leading research institute with a focus on valuable and responsible AI application.

Ron Bodkin joins Vector effective September 8, 2020.


Standardized protocols are key to the responsible deployment of language models

The known unknowns: Vector researcher Geoff Pleiss digs deep into uncertainty to make ML models more accurate

A man looks at a white board with red formulas on it
Trustworthy AI

How to safely implement AI systems