I help teams and organizations research, develop, and deploy AI systems, with a core focus on Large Language Models (LLMs) and Natural Language Processing (NLP). I’ve founded and lead high-performing NLP teams in both early-startup and enterprise contexts. As a leader and individual contributor, I’ve shipped a wide range of ML systems, with a primary focus on putting large language models in production. I thrive under pressure and am driven by a love for building innovative technology that solves hard problems.
Industry & Academic Experience
Currently, I’m an Associate Director of Machine Learning at S&P Global, where I lead a full-stack NLP/ML/Search team in the Engineering Solutions Division. My team is part of a stealth product org that’s developing an AI-powered SaaS platform to help engineers and scientists solve incredibly difficult problems in the information and decision-making space.
Prior to S&P Global, I was a founding member of the AI Team at Mantium, a startup dedicated to putting the power of Large Language Models and generative AI in the hands of non-technical users. Before joining Mantium, I was a Computational Social Scientist in the Machine Intelligence Research and Applications group at Ntrepid, a SaaS organization in Northern Virginia that provides products and services in the cyber-security and AI space. At Ntrepid, I developed and integrated Transformer-based zero-shot inference systems into our product prototypes before “zero-shot” was a buzzword.
Joining Ntrepid was a milestone for me, as it marked my departure from academia. I’d spent the six years prior working at the intersection of statistics, ML, NLP, and social science. I completed a PhD in the Computational Social Science Lab at the University of Southern California and a subsequent Postdoc at Northwestern’s Kellogg School of Management. I helped pioneer the use of modern ML and NLP methods in the study of human social behavior, and my research was published in journals like Nature Communications, Nature Human Behavior, and the Proceedings of the National Academy of Sciences (PNAS). However, as time passed, I became more and more certain that academia wasn’t a great fit for me. I’d always loved building things, and I found that I was happiest when I was working on something that I knew others would use.
So, I took the plunge and joined industry. And I’ve never looked back 😊
Blog
I’ve worked on many aspects of productionizing LLMs and I continue to be surprised by how hard it is. Even with a proliferation of tools that ease parts of the process, it remains difficult to stitch everything together in a way that satisfies performance and reliability constraints. So, I’m launching a blog that will focus on the entire process, from prototyping through deployment!
The blog is a gift to my future-self (and others, if they find some use in it) that aims to treat some of the primary pain points that we encounter when prototyping, productionizing, and deploying ML systems that rely on LLMs. It will focus on deep-dives with comprehensive code and reproducible repositories that can be used as patterns for solutions to common problems.
Consulting and Advisement
I consult on architecture, AI strategy, and end-to-end projects. If you’re interested in working together, please feel free to get in touch!