ABOUT ME

I'm currently a Machine Learning Instructor at Inspirit AI. I am also a Satellites co-chair for the International Society of Music Information Retrieval (ISMIR) 2024. Previously I was a machine learning engineer at companies like Arize AI and World Resources Institute. Read more about my work below.

Work

Machine Learning Instructor @ Inspirit AI (2022 - Present)

  • Taught 600+ students concepts in AI/ML, including using language models, CNNs, BERT, and more
  • Mentored 50+ students on advanced ML projects such as forest fire prediction using visual transformers
  • My students have been accepted to competitive journals such as JSR or have been accepted into internships

AI Engineering @ Arize AI (Summer 2023)

  • Contributed to LLM-based retrieval system evaluation which is now used in Arize Phoenix, 3k+ Github stars
  • Achieved ~70% chatbot context retrieval accuracy (precision, recall) via tuning
  • Implemented state of the art methods such as RAG, Cohere reranking, and HyDE to improve chatbots' response quality

Machine Learning Engineer, Software Engineer @ Tagg (2021-2022)

  • Designed and implemented a recommendation system on AWS EC2 on 3500+ posts for 1500+ users
  • Achieved an average nDCG score of 65-70% for the recommendation system
  • Trained an Interest classifier to classify users’ posts based on their interests using InceptionV3
  • Developed front-end components in React Native

Software Engineer Intern @ Nikira Labs (2020-2021)

  • Reduced code runtime by 56% using efficient data structures and vectorization
  • Recasted a 4k+ lines of code Python 2 codebase to Python 3 using Pandas and Numpy

Data Science and Machine Learning Intern @ World Resources Institute (Summer 2019)

  • Developed a deep learning regression model using Keras to predict CO2 emissions of a power plant, achieving an average coefficient of determination of 0.97 and 13% error rate on the validation set
  • Designed and implemented automated web crawlers to extract 30,000+ data points using the Beautiful Soup

Server Support Engineer @ Stanford University (2016 - 2019)

  • Maintained software infrastructure for the Center for Turbulence Research at Stanford. Presented comparisons of program execution times to HPC director.

Teaching Assistant for ME 344 Introduction to High Performance Computing @ Stanford (Summer 2017)

  • Led 90+ students with computer cluster building projects in an 8 week long course and teaching guide for Intel's high performance computing material. Taught practical topics in high performance computing.

Academics

Stanford University

B.S. in Computer Science, Artificial Intelligence Track

Courses: Principles of Artificial Intelligence, Applied Machine Learning, Deep Learning, Machine Learning, Computer Vision, Natural Language Understanding, Reinforcement Learning, Research in Computer Generated Music

University of Chicago

High School Dual Enrollment Program

Courses: Linear Algebra, Calculus 3, Statistics 1

Projects

Chromesthesia (AGI House Hackathon, 1st Place)

Developed an audio-to-video model using stable diffusion, GPT 4, ImageBind, and inference on H100 GPUs to generate a music video in under 5 minutes from just an audio source

ChatGPT Observe Plugin (Cerebral Valley Hackathon, Finalist)

Deployed a ChatGPT Plugin with early access to plugin APIs. Observe Plugin allows users to link a dataset via a URL and ChatGPT will be able to reason about it

Course Deploy (Cerebral Valley Hackathon, Finalist)

Built a website that dynamically generates personalized homework for students. Built endpoints to GPT-4 API for generation, including code generation and feedback conversations

Volunteering

Satellite Events Co-chair for International Society for Music Information Retrieval (ISMIR) 2024, coordinated venues for conference events and a hackathon