Michelle Gregory

Senior Vice President of Data Science @[24]7.ai

San Francisco, CA, US
MOBILE NUMBERS
+17•••••••33

Signup · Get unlimited contacts

WORK HISTORY

Aug 2020 — Present

Senior Vice President of Data Science @[24]7.ai

View department →

CORPORATE PROFILE: A customer service company with employees. RESPONSIBILITY PROFILE: I report to the CEO with shared responsibilities for team budgets and shared company revenue goals managing 45 employees with five direct. I am responsible for defining the AI roadmap for conversational AI, maintaining a world-class AI solutions team, leading LLM use cases, and acting as the company\'s public face for AI. CUSTOMER ENGAGEMENT • The company required an AI leader for a successful IPO acquisition. I developed an AI roadmap and partners, libraries, and tools, then enhanced the system, created value propositions, and developed communication metrics. This strategy was declared Best Practice, effectively completed data collection and cloud implementation, improved customer satisfaction, and heightened brand awareness. ML OPERATIONAL OPTIMIZATION / DATA MIGRATION • To reduce processing costs, the COO charged me to close data centers, refocus GCP pipelines, and transfer products. Within 12 weeks, formed a strategy and gained approval, then migrated AI models from the data centers, oversaw platform integration, tracked incident reports, ensured algorithm efficacy, and navigated buy-in from teams and clients. This strategy was declared SOP, as it reduced processing costs by $600k while eliminating redundant data centers. AI / MACHINE LEARNING • Personally discovered an opportunity to use a large language processing model to improve business efficiencies. Over 8 months, analyzed research papers on the topic, devised a novel student/teacher methodology, assessed security limitations, and resized the model to fit business needs. These efforts improved intent prediction, led to enhanced containment, and reduced the model size by 13% without declining results; as a result, the model’s usage became an SOP. The model was developed using GPT2 prior to the release of chatGPT3. Since then, we have been working to integrate Generative LLMs into our production workflows.

EDUCATION

1995 — 2001

University of Colorado at Boulder

PhD

This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.