Bo-Wei Chang
CS+DS @ UC BerkeleySan Francisco, CA

ABOUT

I'm Bo-Wei(Warren) Chang, studying computer science and data science at UC Berkeley. I've always been very interested about technology and its potential to make this world a better place. Whether it's developing innovative software solutions or exploring the potential of AI and machine learning, I am driven by a deep passion to use technology as a force for good.

Outside of school and work, I really enjoy playing golf and video games. Golf is a sport that challenges me to think strategically and stay calm under pressure, which are skills that translate well into my work and school. Video games, on the other hand, fuel my creativity and problem-solving skills. Whether I'm on the golf course or immersed in a game, I'm always looking for ways to push myself, learn new things, and have fun along the way!

Experience

TikTok

Incoming Software Engineer Intern

Fall 2026
TikTok logo

Tech Stack

PGC AI Platform Team

Salesforce

Incoming Software Engineer Intern

Summer 2026
Salesforce logo

Tech Stack

Sales Cloud

TSMC

Software Engineer Intern

06/2025 - 08/2025
TSMC logo

Tech Stack

Python · LangChain · LangGraph · Docker

Designed and implemented a Retrieval-Augmented Generation (RAG) pipeline using LangGraph to translate Oracle SQL queries into Trino SQL, enabling seamless migration from a traditional Oracle-based data lakehouse to a Trino-powered architecture. Developed error-handling mechanisms and evaluation metrics to assess translation quality, achieving 97% execution accuracy and 93% result accuracy, significantly reducing manual query rewriting effort during migration.

Lafayette Square

Project Manager/Tech Lead

01/2025 - 05/2025
Lafayette Square logo

Tech Stack

React · Node.js · Notion · GitLab · AWS · CI/CD

Led the development of a customizable report generation platform used for generating ESOP insights. Scoped UI and backend architecture, implemented a CI/CD pipeline to transform Snowflake-driven API endpoints into AWS Lambda functions, and configured API Gateway routing. Managed code quality and collaboration through GitLab workflows for efficient deployment.

Curiocity

Software Engineer

08/2024 - 12/2024
Curiocity logo

Tech Stack

Typescript · React · Node.js · LlamaParse · AWS

Contributed to the development of a document editing platform that supports secure file upload, parsing, and user authentication. Implemented REST APIs with MongoDB and S3, integrated Google OAuth for user access, and built a file ingestion pipeline using LlamaParse to extract structured data from formats like PDF, PPTX, and HTML.

TeachShare

Software Engineer

06/2024 - 12/2024
TeachShare logo

Tech Stack

Python · React · Typescript · NLTK · Javascript

Enhanced a worksheet creation tool by implementing a PDF parsing pipeline using PyPDF2, PDFMiner, and NLTK. Integrated Polotno SDK into the React-based frontend to enable dynamic editing, and added support for editable imports from Google Slides and PDFs via Adobe Extract API. Built automated publishing pipelines using social media APIs.

Wing Assistant

Software Engineer Intern

05/2024 - 08/2024
Wing Assistant logo

Tech Stack

Python · VoyageAI · LangChain · Pinecone

Built a resume intelligence system integrating RAG pipelines and vector search. Retrieved and preprocessed resume data from Lever, embedded using Voyage AI, and stored in Pinecone for semantic search. Enabled ranked candidate matching with justification, improving accuracy and efficiency in hiring workflows.

Ochy

Software Engineer Intern

01/2024 - 05/2024
Ochy logo

Tech Stack

Typescript · React · GCP · Pub/Sub · React

Developed a personalized video generation system using React, Node.js, and Remotion. Parsed structured JSON feedback to dynamically render highlight videos, and deployed a real-time video streaming pipeline using Google Pub/Sub to support low-latency content delivery.

06/2023 — 08/2023
Association for Computing Machinery(ACM) logo

Tech Stack

Python · PyTorch · Tensorflow · Roboflow

Worked on training custom datasets for traffic signal recognition, manually labeling over 4000 images using Roboflow. Implemented the model using YOLOv7, achieving an accuracy rate of 88.15%.