Backend-focused Software Engineer who creates great user-facing products and infrastructures. Experienced with building LLM-powered tools and pipelines. Excels at LLM, Java, Python, Trino, AWS, Google technologies, Jenkins, learning new technologies and combining them with the current knowledge base.

Experience

2022 — NowGoogleSenior Software Engineer

2022 — Now

Irvine, California, United States

Built a RAG-powered Chatbot for Google Sellers to quickly lookup both internal and external knowledge base.

Designed and developed an automatic data serving pipeline as a centralized gateway to serve critical business data across Google PAs as views and materialized views.

Designed and developed a Semantic Cache with vector embeddings for LLM requests.

Designed and developed a pipeline to extract snippets from documents using LLMs, as well as implemented a semantic retrieval engine for a user-facing chat bot.

2021 — 2022TuSimpleSoftware Engineer - Data Infrastructure

2021 — 2022

San Diego, California, United States

● Designed and developed a benchmark framework for Trino to evaluate its performance in terms of metrics,

comparison with underlying storage sources, and comparison with AWS Athena.

● Utilized pyspark to inject benchmark result metrics to TuSimple Data Lake on AWS S3 in Parquet format.

● Configured an AWS Glue crawler to crawl S3 benchmark metrics, and update the corresponding Hive table.

● Integrated Looker as a dashboard to visualize benchmark metrics for business intelligence and analytics.

● Supported other teams to onboard our consumption pipeline of storage (Hive, PostgreSQL, MongoDB), query

engine (Trino), and dashboard (Looker).

● Utilized pytest-benchmark and google/benchmark as the Benchmark frameworks

2020 — 2020IntuitSoftware Engineer - Platform Infrastructure

2020 — 2020

Mountain View, California, United States

● Designed and developed a Python CLI tool to identify system resource wastage in Kubernetes Pods and AWS

EC2 instances, and proposed actions to reduce wastages.

● Collected data from Prometheus, Wavefront, and IKSM using API Clients and RESTful APIs.

● Integrated Google Sheet using Google APIs to generate online report sheets with shareable links.

● Implemented calculation algorithms to convert wastage into numeric cost.

● Applied Bin-Packing Algorithm in estimating number of saved AWS nodes and budget.

● Integrated new Java SpringMVC to resolve Java Bottleneck by saving 60% on CPU and 70% on Memory.

● Saved an average annual cost of around $1M across 200+ AWS Kubernetes clusters.

2018 — 2019DecorMattersSoftware Engineer - Full Stack

2018 — 2019

Milpitas, California, United States

● Designed and developed a dashboard for DecorMatters and business partners to visualize statistics of products.

● Utilized React and Ant Design library to implement a single-page responsive dashboard with user-friendly modern UI designs.

● Used BizCharts as a data visualization tool to create intuitive graphs on statistics.

● Used NoSQL database (MongoDB) to manage account, product, and statistics data.

● Created an admin dashboard for administrators to CRUD on the database without query commands.

● Implemented checkers and converters to validate CSV and JSON files.

● Created a Node.js backend server to handle HTTP requests and responses.

Education

California State Polytechnic University-Pomona

Experience

Education

Master of Science - MS