StroolTrading – Financial AI Training, Data Annotation & Model Evaluation
I contributed to the StroolTrading project by designing, annotating, and evaluating financial text and code-based data used to train an AI-powered trading assistant. Key responsibilities included: Classifying financial events and market conditions Creating and evaluating prompt–response pairs (SFT) for LLM training Summarizing market insights and validating AI-generated financial reasoning Annotating code-related data, including trading logic, technical indicators, and backtesting rules Evaluating model outputs for accuracy, coherence, robustness, and financial realism The project required a strong understanding of financial terminology, algorithmic trading concepts, and time-series logic, as well as high attention to detail and consistency across annotations.