Chris(Haipeng) Wu Resume PDF

Chris(Haipeng) Wu Los Angeles, CA | (323) 719-3730| [email protected] | www.linkedin.com/in/haipeng-wu EDUCATION University of Southern California Los Angeles, CA Bachelor of Science，Computer Science (GPA: 3.89/4)...

Chris(Haipeng) Wu Los Angeles, CA | (323) 719-3730| [email protected] | www.linkedin.com/in/haipeng-wu EDUCATION University of Southern California Los Angeles, CA Bachelor of Science，Computer Science (GPA: 3.89/4) August 2023 – May 2027 Relevant Coursework: Introduction to C++, Data Structures and algorithms, Discrete Mathematics, Linear Algebra, Probability Theory, Multivariable calculus SKILLS Programming Languages: C++, Python, JavaScript Frameworks/Libraries: AWS, Embedded JavaScript, Express.js, Node.js, React.js, Next.js, Git, Pytorch, scikit-learn WORK EXPERIENCE Full Stack Developer | Health Systems Engineering Lab Los Angeles, CA, August 2024 - Present Enhanced a patient decision aid website with 100+ users, boosting user engagement by 10% based on subsequent interviews Added a user query suection, integrating webkitSpeechRecognition API for voice input and utilized Nodemailer Enhanced UI/UX using JavaScript and EJS based on user interviews by improving CSS and page layout Redeployed website on AWS Elastic Beanstalk and integrated MongoDB, streamlining deployment with CodePipeline Strengthened codebase security by managing sensitive credentials with dotenv, replacing hardcoded passwords Product Analyst Intern | Softbank China Venture Capital Shanghai, China, June 2024 - August 2024 L  Investigated 10+ pitches of frontier startups in the field of robotics and AI infrastructure  Conducted product review and market research in AI coding and AI companion agents  Conduct 4 visits and prepare 20 page +research reports for AI application-related portfolio companies PERSONAL PROJECTS Full Stack Developer | QuizAI Chrome Extension Los Angeles, CA, August 2024 - Present  Developed QuizAI, a full-stack Chrome Extension using React.js, Node.js, and MongoDB to generate AI-powered quizzes  Integrated Google Gemini API to generate quizzes, parsing AI responses into structured JSON for frontend consumption  Implemented secure user authentication and authorization using JWT, handling registration and login with Express.js and Mongoose, , handling asynchronous operations and data fetching with fetch API  Utilized React Hooks and React Router for state management and client-side routing, enhancing user experience  Employed Vite for optimized development and build processes, improving application performance and loading times  Connected to MongoDB using Mongoose, performing CRUD operations and managing data schemas efficiently Back-end Developer | E-commerce Database & Search Engine System Los Angeles, CA, August 2024 - Present Developed a comprehensive E-commerce Database Management System using C++ with extensive object-oriented programming (OOP) principles, resulting in 40% faster data retrieval Designed and implemented a custom DataStore class to store and manage product and user data, employing STL containers like vector, set, and deque to efficiently handle dynamic datasets Built custom parsers for data extraction and management by leveraging file I/O and string stream manipulation techniques Ensured modular design with polymorphism and inheritance for product categories, while employing function pointers for customizable SectionParser handling Developer Researcher | Speaker Recognition System Shanghai, China, May 2024 – June 2024 Implemented MFCC-based feature extraction on large-scale LibriSpeech audio datasets using the Librosa library, enabling the transformation of raw audio data into feature vectors for improved speaker classification accuracy Applied a triplet loss function to optimize speaker embeddings, enhancing the model’s ability to differentiate between distinct speakers while minimizing intra-speaker variability, leading to more robust speaker identification Trained a bidirectional LSTM model using PyTorch on GPU to capture temporal dependencies in audio sequences Achieved an Equal Error Rate (EER) of 0.07 at a threshold of 0.588