Software Engineer & Berkeley EECS graduate interested in backend development & machine learning. Add me on LinkedIn or Github!
- Led a large-scale data migration of over 3 million rows from Snowflake to SQL Server by configuring a cron job to enable data synchronization, facilitating OLTP access for product development based on this data
- Managed the experimentation platform and implemented new gRPC endpoints, facilitating A/B testing for feature development and increasing developer productivity by 20%
- Developed the Seller Performance dashboard, a metrics-driven page designed to help sellers improve their ticket purchase rates, with visualizations built in JavaScript and seller data queried through Snowflake
- Alexa Deep Learning Core Team: developing low-latency and memory-optimized inference engines in C++ for both in-cloud and on-device Alexa Speech models, leveraging both generic and specialized chips
- Successfully integrated AWS Neuron Runtime library into Alexa's inference engine, enabling deep learning inference on AWS Inferentia hardware and resulting in a 4x cost reduction compared to NVIDIA CUDA with equivalent latency
- Utilized multithreading on Inferentia hardware to increase throughput by 2x
- Optimized DDSQL, a language modeled after SQL, to comprehensively query over Datadog's logs, metrics, and APM services
- Implemented RBAC (Role-Based Access Control) for a subset of queries, improving overall security for customer data
- Conducted extensive query shadowing for millions of DDSQL queries, resulting in a 50% improvement in query language support
- Successfully optimized query performance for time-intensive queries by 4x
- Assisted 50+ students in weekly lab and discussion sections for CS61C (Computer Architecture) by teaching concepts of C, virtual memory, and caching
- Aided students in office hours with any questions about homework or projects
- Conducted exploratory data analysis on a large dataset of 50,000 songs from the Spotify Web API to extract unique characteristics using Python and data analysis libraries
- Developed and optimized a machine learning model using the k-nearest neighbors algorithm to accurately categorize the emotion of a particular song in a user's listening history
- Implemented an interactive classification feature that provided users with a detailed breakdown of their overall mood
- Implemented a version-control system, similar to Git, that featured 13 commands such as add, commit, merge, init, checkout, and branch
- Designed multiple classes to represent the internal structures during execution and a parallel representation as files to ensure the persistence of the program
- Developed a fully functional Scheme interpreter using Python, which can loop, read, evaluate, and print
- Processed parentheses from Scheme syntax using recursion to evaluate and print functions
Lifting, Football, Basketball, Hiking, Swimming -- I'm a pretty active individual! When I'm not running around, you can find me listening to music or playing GeoGuessr.