San Francisco, California, United States
👋 I've called four countries home: Vietnam, Singapore, Denmark, and the U.S. And each one has shaped how I think, work, and connect with curiosity, creativity, and diligence. Lately, I’ve been geeking on AI prompt engineering, agentic workflows, and backend system design. I tend to explore how language models can think, reason, and act through clever orchestration and infrastructure. Outside of code, you might find me with a philosophy book in hand. One of my all-time favorites? When Nietzsche Wept, a thought-provoking mix of psychology, introspection, and a little existential spice. Let’s connect if you’re into building the future of intelligent systems or just want a good book rec!
Software Engineer II • Spearheaded the design and prompt optimization of an agentic-AI Payment Optimization Assistant using Langchain, LLMs, and RAG, enabling autonomous financial decision-making via multi-step reasoning • Developed REST APIs and engineered backend infrastructure for ROI analytics, including schema design for new financial data models such as virtual card rebates and supply chain financing terms. • Integrated ROI services into agentic workflows, enabling real-time response generation tied to financial goals such as cash flow optimization and discount capture. Software Engineer I • Designed and implemented supplier logo recognition for invoices to enhance supplier matching by 18% for an invoice entries automation on the account payables team of Fusion Application • Conducted root-cause analysis across 80+ customer-reported bugs to identify code failure patterns and proposed code fixes and documentation enhancements that reduced support overhead by 15% • Resolved more than 25 customers’ bugs and conducted in-depth analysis on the code base to troubleshoot
• Automated the process of extracting companies’ names and themes from their summary pptx slides using BART-large model fetched from HuggingFace platform and spaCy natural language processing pipelines • Deployed an API utilizing Tornado for the name and theme extracting model and tested it with Postman • Enabled a search engine with ElasticSearch using extracted data for bankers to conveniently look up clients’ profiles
• Acquire production engineering skills such as overseeing the reliability and scalability of services; working hybrid between software & systems engineering • Learn how to administer, configure and upgrade Linux systems, along with the tools and concepts necessary to efficiently build and manage a production Linux infrastructure. • Contribute to open source projects and get consulted by expert mentors on a daily basis
• Investigate the consistency of multiple models of automated voice processing systems (VPS) • Determine the degree to which explanations vary over these VPS models (which have varying structures) and consider the reasons for this variation. • Learn the foundational material necessary to understand the VPS including the classification of graphemes, the LIME explanation method, and the AVP pipeline