Industry and Academia


Transformers & the New AI Compute Pardigm

(Nov 2023) An overview of the universal mathematical and computational theory behind ubiquitous AI Transformer models like BERT, GPT and T5 that dominate state-of-the-art AI research and industry application today. Concludes with future research and industry developments including the new AI Computing Paradigm and enterprise computing ecosystems arising around AI like Semantic Kernel and AutoGen from Microsoft and Open Source alternatives like LangChain, LlamaIndex and Stanford’s DSPy.

AI Integrated into Higher-Ed

(Oct 2023) On-site workshop to strategize on comprehensive integration of AI into an existing program. Based upon the many lessons and best practices learned from building the world’s first human-centered AI curriculum and original project-based research AI lab.

Human-Centered AI Curriculum

AI & Manufacturing

(Sep 2023) Presenting my original works on using LLM to augment and generate synthetic data for manufacturing simulation. NAFEMS is the only worldwide independent association dedicated to engineering modelling, analysis & simulation, across all industries. We are a global community of nearly 25,000 individuals belonging to more than 1,500 organisations, from engineering consultants to major corporations.

AI & Finance

(May 2023) Research, prototype and deliver an end-to-end automated workflow for raw audio to final analyst report generation for VC and private equity investors. Explore dozens of small to midsize LLMs as for greater privacy, flexibility and data security. Compare various traditional NLP, prompt engineering and LLM models for topic identification and summarization. Justify and quantify project ROI based upon detailed cost accounting and practical legal license restrictions.

Workflow Automation and Report Generation
Limits of AI Cognition

AI & Intelligence

(May 2023) Panelist on webinar by Mark Daley

In the increasingly digital world, large language models (LLMs) powered by AI are becoming more sophisticated and pervasive. This talk explores the evolution of neural nets from their inception by McCulloch and Pitts, through the early computational perspectives of Putnam and Fodor, to the current LLMs that challenge our understanding of human creativity, intelligence, and identity. It probes Sutton’s “bitter lesson” of the necessity of computation and scrutinizes examples of intelligence, while questioning Aaron Sloman’s concept of “exotic minds.” The conversation culminates in the Singular Intelligence Hypothesis, suggesting that computational rationality could serve as a bridge between human and machine decision-making and behavior.

AI & Healthcare

(Apr 2023) Technical presentation on the history, architecture and diverse medical applications of Large Language Models for AI Experts Network in HealthTech. Presentation into rapidly evolving models, benchmarks and eco-system around LLMs. Give an integrated perspective into current and future technical, legal and business aspects of LLM-assisted medical applications including hallucination, privacy and business risks like intellectual property/licencing. Provide a forward-looking strategic plan with automated agents, tool chain use (LangChain) and Distributed Autonomous Organizations (DAOs).

AI Large Language Models in Healthcare
Al Disruption in Higher Ed

AI Strategies for Higher Ed

(Jan 2023) Give three distinct presentations to the Kenyon College community, communications and admissions offices on the challenges and opportunities of AI in Higher Education. Review our successful human-centered AI curriculum, popular interdiciplinary research, and offer multiple strategies tailored to our institution’s strengths.