
Intern
Infosys · Jan 2024 – Present
- A real-time AI-based fraud detection system leveraging deep learning, business rules, and explainable AI to identify suspicious financial transactions.
- Worked on cleaning and preprocessing large BFSI transaction datasets for model development.
- Contributed to feature engineering, creating meaningful inputs to improve fraud detection accuracy.
- Collaborated during model training where Random Forest and Logistic Regression showed poor performance on imbalanced data.
- Worked with team and mentor guidance to transition to LSTM + Transformer models, improving prediction performance.
- Assisted in evaluating model performance, resolving issues, and refining the ML pipeline.

