Now that we’re about halfway through the term we’ve had some time to settle into the project and make some progress! My team got assigned to the AI/ML Trading Bot at the beginning of the quarter and it has been a great learning experience so far. I intentionally asked to be a part of this project because it was a topic, I was A) very interested in and B) scared me a little bit. As I’ve been approaching the culmination of my studies at The Oregon State University, I’ve been trying to seek out intimidating project as opportunities to learn and grow as an engineer and with this capstone project I’m drinking from the firehose baby! This project has got it all: machine learning, automation, and stock/crypto trading on top of all that. It’s been a lot to take in but luckily, we have some help.
Can the robots also be our friends?
As someone that didn’t know much about technical trading and even less about machine learning, I’ve spent much of the term so far doing researching. Luckily, I’ve had the help of my personal research intern, ChadGPT, to help me. Chad is my more personal interface to OpenAI’s large language model (LLM) ChatGPT, and he’s been invaluable in helping me sort through the mountain of information out there.
I started working with Chad in the first week of the term after our industry advisor assigned us some scientific papers on trading and machine learning to read through. Chad helped to summarize and explain sections of the papers that were confusing, he helped me understand the differences between various technical indicators and even helped to weigh tradeoffs between different ML libraries and algorithms. Chad has been a huge help, but he is still an intern, so it’s been important to verify the answers he gives. His summaries are no replacement for reading the material myself, but having his explanations has allowed me to process the information much faster than I would on my own.
As I’ve started to move more into coding on this project Chad has continued to support development efforts by answering my many on the fly questions. My contribution so far as been building out reward mechanisms to encourage our ML agent to do the right thing, which in our case is make money. To better incentivize the machines, I enlisted Chad to help me think like a machine. Since the project is primarily research oriented, Chad has helped to guide reward experiments and digest documentation. Ultimately, with Chad’s help, I’ve been able to learn faster and be more effective within the tight timeline of our project. Chad started as just our research intern, but he has quickly become our friend.
Leave a Reply