Author Interview with Yiqiao Yin
Author of the books: Notes on Agent-based Applications
1. Please introduce yourself. What would you like your reader to know about you?
-I come from Shenzhen, China, from a very conventional Chinese family. My parents both work full time and I sometimes get takeouts after school. I came to the United States in 2010 for college but I did not study computer science in undergrad. Back then I wanted to do finance so that's what I did. I finished with a Mathematics degree and decided to work on Wall Street as a trader. Shortly after that, I realized my potential to use code to maneuver large-scale cross-sectional financial datasets which gave me the inspiration of doing data science as a career. After 10+ years of working my way from Data Scientist, Researcher, Machine Learning Engineer, Tech Lead and now a Principle AI Engineer at a S&P 500 company, I am grateful to every step of my career and those who I worked for and with. I am excited to share some of the most advanced lessons and notes I learned in Machine Learning, Artificial Intelligence, and GenAI with you.
2. What is your inspiration/motivation?
-Machine has been and always will be a logical extension of human will. That has been my faith and will be an unsaid theme throughout my work and, of course, all my books as well.
3. How long did it take to complete your book from the idea to publication?
-The first edition took me a year to consolidate. As the title of the book describes, it is indeed a series of notes I have taken in the past and it is something I will iterate on. As time moves on and more progress has been made, I felt the need to finish the second edition which takes a little shorter than a year. Together, I proudly present my notes on agent-based applications that compile some of the most advanced practices using Large Language Models (LLMs) and GenAI tools.
4. What's the main message and idea of your book "Notes on Agent-based Applications(2nd Edition)"?
-The novelty of the book lies in its laser focus on a proprietary package called the WYN-Agent package, a Python-based, pip-installable tool that enables users to interact with virtually any computer analysis and run simulations using Python with a natural language interface. An upgrade to WYN-Agent inspired the development of WYN-Agent-X, which builds upon the foundational concept of WYN-Agent but is now enhanced with mouse-moving capabilities, allowing users to control their laptops through a natural language interface.
5. What was the most unexpected conclusion you came up with while preparing "Notes on Agent-based Applications(2nd Edition)"?
-It has come to my attention that agents can now handle most basic tasks. However, sequential and hierarchical planning still remains a human responsibility. I don’t see an easy way to develop a generalized planning strategy that can be delegated to agents. This challenge will take time to resolve, but it is not entirely impossible. For example, a book demonstrated a procedure in which a chatbot controlled a mouse cursor to purchase cat food on an Amazon shopping site. However, critical steps, such as deciding which cat food to buy, still require human intervention. It remains difficult for algorithms to determine whether I am tired and just want to complete the purchase quickly or whether I am feeling energetic and willing to shop around and research a cheaper option.
6. How would you describe your publishing experience with Eliva Press in a few words?
-I have a very good relationship with Eliva. I have worked with Olga many times and she has been very helpful to guide me through the publication process. I am looking to continue to work with her in the future.