Inspiration
We wanted to tackle a AI-driven real-time trading — building a bot that reacts to news, trades fast, and handles risk like a real market maker.
What It Does?
Market-makes continuously on all primary stocks Detects trading oportunity depending on sentiment from social-media posts using LLM MM and aggressive sentiment trades
How We Built It Python + Optibook exchange client
Challenges
Avoiding breaches of position limits Handling rapid clustered news events Keeping MM and sentiment from interfering with each other
Accomplishments
Robust MM–sentiment coordination Stable quoting, safe risk handling, and clean, explainable logic A system that feels like a realistic trading engine Implement basic dual-listing arbitrage engine
What We Learned
Risk controls matter more than clever strategy NLP is powerful but must be tempered with confidence, cooldowns, and exits Real-time trading is mostly engineering
What’s Next
Fine tuning :)
Built With
Python • Optibook • Transformers • PyTorch and ❤
Log in or sign up for Devpost to join the conversation.