Dr.K.Shiva Keshav Reddy,L.Yamuna,K.Narendra 02 05(1) Risk management is at the heart of investment banking where institutions face a spectrum of risks including market risk credit risk liquidity risk and operational risk Traditional risk management relies heavily on Value at Risk VaR stress testing scenario analysis and expert judgment to measure monitor and mitigate these exposures However these methods often assume linear relationships and static volatility which may fail to capture realworld complexities and sudden market shocksThis study combines conventional risk management frameworks with advanced analytics Using Machine Learning models like Random Forest and XGBoost we aim to detect hidden nonlinear risk factors and predict potential losses more accurately Additionally Deep Learning models such as LSTM networks are employed to forecast timeseries risk metrics like daily VaR and liquidity ratios effectively capturing volatility clustering and longterm dependencies in financial dataBy integrating ML and DL into risk analysis the study demonstrates significant improvements in predictive accuracy and early warning capabilities offering investment banks a more dynamic and datadriven approach to risk management Copyright (c) Journal of Science Engineering Technology and Management Science. All rights reserved 2026