Engineering Intelligent Systems for Finance & Electrical Infrastructure.

Architected a hybrid deterministic-AI platform for autonomous electrical grid resilience. Integrated a LangGraph multi-agent pipeline with a pandapower Digital Twin and Soft Actor-Critic (SAC) reinforcement learning policies for active power redispatch. Engineered a predictive LightGBM forecasting layer and a strict physics validation firewall that successfully intercepted 100% of LLM-induced mathematical hallucinations to optimize N-1 contingencies.

Designed a systemic market stress forecasting system utilizing a 41-feature hybrid pipeline of quantitative indicators and FinBERT sentiment analysis. Engineered a cost-sensitive XGBoost ensemble model, accompanied by a real-time inference dashboard featuring SHAP explainability.

Architected a quantitative finance framework to mathematically balance capital appreciation, tail risk (CVaR), and Environmental, Social, and Governance (ESG) mandates. Implemented the NSGA-II evolutionary algorithm paired with an XGBoost classifier to dynamically optimize asset weights.

AI-driven health monitoring and Remaining Useful Life (RUL) prediction framework for critical high-voltage substation nodes. Directing a dual-model ML engine utilizing Random Forest ensembles for discrete fault classification and LSTM networks for continuous degradation modeling, fused into a native Streamlit HMI.

Developing an intelligent generative AI advisory platform to autonomously process complex admission queries and guide academic workflows. The architecture focuses on implementing robust RAG evaluation metrics, fine-tuning models for structured institutional data, and establishing deep system observability to ensure production-grade reliability.