Enabling enterprises to accurately and easily connect internal data feeds to large language model workflows.

Teaching the Machine Learning Foundations track with Mike Wu, Julie Kallini, Kevin Wu, and Eric Wu. I teach two 4 week project-driven online courses covering ML modeling and best practices for any scale of ML project. Introduction to Applied ML is for new practitioners, and Data-Centric Deep Learning covers more advanced approaches and tools to get deep learning working on real problems.https://uplimit.com/track/ml-foundations

Teaching Spoken Language Processing (CS224S)Course covers traditional and deep learning approaches to building conversational agents and spoken language understanding systemsLecture slides and course details:http://cs224s.stanford.edu

Special Projects Group.Team lead building algorithmic approaches and tools for large-scale, data-centric deep learning. My teams are responsible for producing high quality labeled data with a mixture of human and machine annotation, along with data-centric tooling to address long-tail deep learning performance challenges for robot autonomy.

(Acquired by Parexel) Roam empowers healthcare companies to derive structured meaning from language data to improve healthcare for all.Defined product and technology strategy. Built and led cross functional engineering teams working on a platform for healthcare natural language processing. Led technical sales efforts and requirements gathering from customers.

Building neural network approaches to spoken language understanding.Acquired by Microsoft in 2018

Dissertation: Deep Neural Networks in Speech RecognitionAdvisor: Andrew NgSecondary Advisor: Dan JurafskySee my research website for details.

Worked closely with engineering team to architect approaches to speech recognition and spoken language understanding to enable the wit.ai platform for developing spoken language understanding models.Wit.ai was acquired by Facebook in 2015, and went on to form the foundation for Facebook's M Messenger agent, as well as other spoken language understanding projects.

Building machine learning tools for massive online education

Personalized and predictive navigation solutions

Worked with Brian Kingsbury on the speech research team. Developed novel deep neural network approaches to acoustic modeling for multiple languages, and with limited amounts of training data.

Worked with Dong Yu and Li Deng on the Microsoft Research Speech Team. Developed deep recurrent neural network acoustic models for conversational telephone speech

Developed technique for one-shot learning with Bayesian networks. Resulted in a paper presented at Cognitive Science Conference in 2009

Assisted Prof. Tai Sing Lee teaching computer vision in spring 08 and spring 09.

Advised by Dr. J. Andrew Bagnell. Developed particle filtering tools for GPS data. Led a small software development team to create personal navigation device software

Advised by Dr. Chris Atkeson and Dr. J. Andrew Bagnell. Developed robotic simulation software. Developed Kalman filter for sensor integration. Assisted graduate students in preparing software for monthly evaluations

Lotus Notes database development