A wide range of experience of full-stack web and mobile development in Java, JavaScript, Python, Go, Spring, React, Vue and React Native.
Experience
2024 — Now
2024 — Now
Cupertino, CA
• Architected a low-latency AI agent platform using React.js, FastAPI/Python and Server-Sent Events (SSE) to stream LLM responses in real time, powering automated research and report generation workflows.
• Built a rich-text AI report editor with React, TypeScript, Tailwind, and Lexical, rendering structured outputs with interactive charts using Chart.js and D3.js.
• Implemented Redis-backed session and report caching to reduce prompt payload size and optimize context reuse, cutting LLM API costs by ~35%.
• Designed a state-machine SSE stream parser with prefix matching, partial-tag carryover, and schema-validated structured block emission, improving streaming stability and reducing malformed structured-output errors.
• Introduced an Indexed DB-backed local queue for high-frequency document edits, batching rapid updates with debounced backend sync to reduce write amplification; additionally cached recent report state and agent conversations locally to enable instant reloading and session recovery.
• Served images and media assets via CDN to significantly reduce frontend load latency and improve dashboard performance.
• Built a distributed scheduling and media generation system using AWS EventBridge Scheduler, Lambda, DynamoDB, and Celery workers, powering automated jobs and scalable image/video generation pipelines with real-time progress updates via WebSockets.
• Implemented Stripe-based subscription and payment infrastructure with Node.js/Express and DynamoDB, supporting webhooks, proration, promotion codes, and refunds.
• Integrated PostHog to track user behavior, feature usage, and workflow engagement, enabling data-driven product iteration.
2025 — 2026
2025 — 2026
San Francisco Bay Area
• Designed and implemented a DAG-based workflow execution engine supporting dependency-aware scheduling and configurable node concurrency, enabling parallel execution of image and video generation pipelines.
• Implemented input-hash node caching and snapshot-based partial recomputation, allowing workflows to reuse previous results and rerun only modified nodes instead of the entire pipeline.
• Developed an asynchronous AI inference pipeline using Celery workers and webhook callbacks, integrating AIHub to handle long-running image and video generation tasks.
• Built a visual AI workflow builder using React.js and ReactFlow, allowing users to design custom pipelines for automated image and video generation.
• Optimized canvas performance by implementing debounced snapshot persistence with AbortController cancellation, preventing redundant save requests during node drag operations and improving editor responsiveness.
2024 — 2024
2024 — 2024
California, United States
• Developed frontend using Next.js, Tailwind CSS, TypeScript and Zustand, integrated a Notion-like web editor built with Tiptap, driving an increase of 1000+ new users.
• Connected the StealthGPT API to humanize AI-generated articles and integrated the Copyleaks API to check the plagiarism and AI-generation percentages for articles.
• Automated detection of unused images and variables in files by writing Python Automation scripts.
• Hoste a FastAPI/Python and PostgreSQL backend service using Prisma ORM on AWS, enabling real-time article updates.
2024 — 2024
2024 — 2024
San Diego Metropolitan Area
• Designed and implemented a login system using Firebase Authentication SDK and ContextAPI, adding Google/LinkedIn login options to improve user accessibility.
• Integrated Mixpanel to monitor click-through rates and user dwell time on each interface and used analytics to identify key drop-off points and guided a complete app redesign.
• Reconstructed 100+ pages and launched a new "Story Library" feature where users can share, like and save stories with React Native and Firebase, attracting 3000+ new users and improving engagement metrics.
• Built a Node.js/Express.js server with Redis caching, reducing response latency by 40%.
• Developed and deployed RESTful APIs using FastAPI for multiple ML/AI models, ensuring efficient and reliable performance on Google Cloud Platform through load balancing and API gateway setup.
2024 — 2024
2024 — 2024
Baltimore, Maryland, United States
• Developed RESTful APIs from scratch using Node.js and Express and integrated with Name.com APIs for domain reselling.
• Created domain management dashboard with Next.js, TypeScript, Redux Toolkit and Tailwind CSS.
• Integrated Stripe SDK and PaymentElement to facilitate secure transactions on the platform.
• Optimized frontend search functionality by implementing AbortController to cancel redundant HTTP requests.
Education
Washington University in St. Louis
Master of Science - MS
The Chinese University of Hong Kong, Shenzhen 香港中文大学(深圳)