Inspiration
We were inspired to build this project by our experience interacting with the conventional brokerage apps that provide a very limited toolkit to their customers and lag quite substantially in the development of their software base
What it does
TradeSphere AI is a multipurpose platform that can be used to backtest trading strategies in a variety of ways while offering abstraction to the end user from the pure code which is enabled by our API
How we built it
We started with foolproof financial algorithms and data structures to enable abstraction both for us and for the AI chatbot. We then used to build our portfolio simulation environment and ultimately the front-end.
Challenges we ran into
Initially the idea was to abstract from Python to simplified DSL-like code that the user can write which is then translated straight to Python. However, we soon realized that it would requite a much more significant amount of effort than we anticipated so for the purposes of the hackathon we opted to simulate a modest portion of our intended functionality via GUI
Accomplishments that we're proud of
Creating a working AI chatbot was a great achievement for me and Ron, and we were happy to see it learn on some our prompts to give better advice to the user (although there is still much to be desired)
What we learned
We learned quite a lot about what it means to be a full stack developer and how to connect front-end applications to your backend. We learned to access the external LLM APIs and use them appropriately
What's next for TradeSphere AI
There is a lot of potential for expansion of functionality: we can add more informative graphs, ability to buy and sell based on RSI, Beta, Moving Average, ability to test on real time price data, etc.
Log in or sign up for Devpost to join the conversation.