Wei-En Lin (林煒恩)

Collaborate with clients to understand business needs and propose tailored GenAI solutions that drive innovation and efficiency. Lead PoC initiatives by designing, developing, and validating GenAI prototypes to demonstrate real-world business value.

Developed AI ChatBots using AWS components and LLMs to meet company requirements. Utilized prompt engineering techniques to enhance performance and fulfill client needs.

• Developed the collection scorecard model with PostgreSQL, Python, and Airflow to determine the willingness of customers to pay back the loan after defaulting. Minimize the necessity for debt collection calls by reducing over 10,000 phone calls. Reduced SQL ETL processing time from 48 hours to 3 hours. • Implemented advanced prompting techniques within a hierarchical structure to extract valuable insights from customer service text files, significantly improving customer experience and uncovering new business opportunities. Implemented a robust error information regeneration mechanism and conducted thorough content checks for enhanced accuracy. • Conducted in-depth research on sentence embeddings, and implemented sentence embeddings in antimoney laundering applications, resulting in a significant 40% increase in precision. • Organized four internal GenAI workshops with over 300 colleagues through engaging presentations, and created the prompting materials, including examples tailored for marketing, code generation, image prompts, and more.

* Coordinated with various departments and assisted in managing key accounts for over 20 clients by conducting data analysis of sales revenue. * Conducted market research on cosmetic packaging materials, identifying pain points and opportunities for growth. Presented research findings at the Asia Pacific Specialty Materials Conference, sharing industry insights with colleagues from around the world.

Conducted research on optimizing chemical separation processes using optimization algorithms, with the goal of minimizing costs while maintaining efficiency. Utilized regression model to predict optimal separation methods for new systems.