Mohsen Moradi Moghadam

Data Scientist @DTE Energy

Detroit, MI, US

Developed and fine-tuned GPT/DeBERTaV3 models in Databricks using MLOps for hierarchical multi-level classification of customer complaints and service requests. Achieved 0.81 weighted F1. Applied advanced expertise in prompt engineering, few-shot learning, RAG, and LangChain to design and deploy GPT-based pipelines, significantly improving model performance. Engineered a ReAct-based agent to dynamically choose between RAG and SQL query modes, enabling context-aware generation of analytical and feedback reports. Designed and deployed an ASP.NET chatbot UI with Databricks + API integration, deployed via Azure DevOps. Designed SAP SQL-based benchmarking reports for Billing, Revenue Mng, and Customer Service operations across multiple departments.