Hi, my name is Fisher and I am a Python Developer, Quant Developer, and AI Engineer. I worked in an AI startup and cryptocurrency exchanges. My expertise is building speech recognition model, text-to-speech model and natural language understanding model in conversational AI and trading system

I have always believed in agent-based AI, which allows you to collaborate with it to solve tasks together. Deep reinforcement learning has been the primary approach for creating interactive AI in recent years. However, training AI agents using this method requires a significant amount of interaction data, which may not be practical for real-world problems.

Now, there is another potential solution: LLMs trained on massive amounts of data. OpenAI's ChatGPT, with its astonishing knowledge, reasoning ability, and understanding of human language, has showcased its immense power to the world since last year. Consequently, I believe it is the opportune moment to embark on building AI agents capable of seamless interaction with humans.