Full CV here.
⌨️ Professional Experience
Zoom
Machine Learning Engineer
Remote, US
June 2022 - Present
- Led the development and evolution of Zoom's "Next Steps" AI—advancing from BERT-based task extraction to LLMs while maintaining continuous quality and performance optimization over multiple years.
- Patented a novel evaluation method for AI-generated action items that improved accuracy and reliability.
- Solved long-context compression challenges by developing chunking and retrieval strategies, enhancing LLM performance for meetings longer than one hour.
- Built "ReScribe," an AI Hackathon finalist project that integrated OCR and LLMs to enhance ASR transcription accuracy.
- Conducted rigorous A/B testing and benchmarking of LLM performance, comparing in-house solutions against GPT-4 and Claude to optimize task generation across features.
- Engineered federated LLM evaluation frameworks using zero-shot ranking and scoring to optimize results.
Neo.Tax
Software Engineer Intern
Palo Alto, CA
June 2021 - September 2021
- Built an interactive devtool from the ground up that automates queries and mutations of live customer data
- Wrote typescript services that interlinked slack webhooks, data on redis cache, live database mutations, more
- Learned and harnessed GCP via rigorous access control in my bot’s BigQuery integration
- Designed and built a Retool data manipulation app equipped for accelerating in-house tax analyst workflow.
Stanford CURIS Computer Science Research
Undergraduate Research Intern
Stanford
June 2020 - September 2020
This past summer I worked under Professor James Landay's lab in the HCI Department to come up with, build, and deliver research that would enable technologies that weren't fused together before.
What I did:
- Developed 2D Depth Sensing by combining Augmented Reality with the always-on nature of monocular cameras
- Created an algorithm that ingests a 3D Space/Pointcloud and returns the location of the Camera’s position
- Utilized cutting edge research and repositories to project a 3D Space from a 2D camera picture
- Built customized ROS tool leveraging latest Simultaneous Localization and Mapping Libraries
- Presented research to faculty and in poster session to the department and public
https://s3-us-west-2.amazonaws.com/secure.notion-static.com/febc50e1-b41b-4e6a-a84c-ee3e1b4edfa7/autotrak_poster_final.pdf
Software Engineer Intern - AI Simulation Team
Portola Valley, CA
June 2017 - June 2018
- Simulated and fine-tuned self-driving car algorithms, testing vehicle AI behavior on game software Unity
- Built fully functional particle-based LIDAR sensor system leveraging Unity engine raycasting
- Interpreted LIDAR sensor readings to analyze decision-making impact of data collection speed vs vehicle speed
- Managed deliverables and feedback in JIRA between self-driving teams and simulation team