China
Private LLM Deployment for Government-Owned Finance Entity Role: Solution Architect (Pre-sales) | Tech Stack:DeepSeek-VL, Alibaba Cloud OCR, Kubernetes, API Gateway - Led the technical design and pre-sales proposal for deploying an enterprise-grade LLM in a fully privatized environment. - Coordinated with model vendors to fine-tune use cases like document summarization, legal policy Q&A, and internal knowledge retrieval. - Delivered PoC that outperformed legacy Q&A systems by 35% in accuracy and response speed. - Worked closely with the client’s cybersecurity team to meet internal audit and data sovereignty standards. AI-Assisted Document Automation for Central State-Owned Ports Group Role: Solution Consultant | Tech Stack: OCR + LLM, RPA, AI Workflow Engine - Designed an AI-based document processing solution integrating OCR and domestic large models to automate contract review and compliance checks. - Enabled real-time extraction and summarization of key clauses from scanned documents with over 90% accuracy. - Integrated digital workers (RPA bots) to trigger automated workflows based on AI outputs, reducing review cycle from days to hours. - Delivered in a hybrid cloud architecture with fallback mechanisms for offline access.
Team Management - AI Startup Team Growth: Scaled team from 4 to 30+, achieving profitability through strategic recruitment and team-building. Led expansion for nearly 10x growth in 12 months. - Recruitment & Culture: Guided new members in technical skills, established workflows, and fostered an open, collaborative environment. Organized knowledge-sharing sessions and discussions. Development Team - User-Facing Website: Built an e-commerce platform with Python & JavaScript, enhancing real-time search and recommendations, increasing user engagement by 25% and satisfaction by 15%. - Machine Learning Pipeline: Led prompt engineering for large language models, optimizing automation and improving model performance by 20%. - Database Optimization: Migrated from TinyDB to MongoDB to support 10x user growth, improving efficiency and scalability. Operations Team - LLM Commercialization: Analyzed user needs, introduced tailored AI solutions, and expanded market opportunities with sustained positive feedback. - Marketing & Engagement: Designed interactive campaigns (React, Node.js), integrating AI-generated voice interactions. Campaigns doubled new customers and boosted sales by 30%. Product Team - Product Research & Roadmap: Led feature research, contributing to three new product developments in six months. Streamlined deployment, cutting production cycles by 40%. - Competitive Analysis: Researched AI competitors to refine commercialization strategies and enhance differentiation.
- Developed frontend and admin pages using Java and TypeScript for an online learning platform, implementing core CRUD (Create, Read, Update, Delete) functionalities. - Built an internal dashboard tool similar to Trello and Jira using the Spring and React frameworks, featuring file upload progress bars, tag management, and more. This tool significantly improved task tracking efficiency for cross-functional teams. - Practiced agile development and continuous integration, including pair programming and daily code reviews. Integrated Jenkins for CI/CD, reducing development cycles by 30% and improving code quality.