Chandu S

About Chandu S

Innovative Generative AI Data Scientist / Sr AI-ML Engineer with 9+ years of IT experience, specializing in LLMs, Generative AI, NLP, and Machine Learning. Proven track record of designing, fine-tuning, and deploying production-grade AI systems using GPT-4, LLaMA (2/3/3.1), Mixtral, Gemini, and Hugging Face models.Hands-on expertise in Retrieval-Augmented Generation (RAG), LangChain, LlamaIndex, and vector databases (Chroma, FAISS, Pinecone), building scalable semantic search and AI-driven applications. Strong background in prompt engineering, parameter-efficient fine-tuning (LoRA, QLoRA), and embedding optimization for domain-specific use cases.Experienced in end-to-end ML/MLOps lifecycle—from data engineering and model development to deployment and monitoring—across AWS, Azure, and GCP using SageMaker, Kubernetes, Kubeflow, and MLflow. Adept at translating complex business problems into impactful AI solutions, collaborating with cross-functional teams, and delivering measurable business outcomes.

Experience

1. Sr Ai Engineer

   Fannie Mae

   May 2023 — Present · WA, US

   I work on analyzing claims and risk-related datasets to identify patterns, anomalies, and high-risk indicators that impact claim outcomes and operational efficiency. I partner with stakeholders to translate business requirements into data-driven solutions, focusing on fraud detection, risk assessment, and predictive modeling.Conduct exploratory data analysis (EDA) to uncover trends, correlations, and outliers.Preprocess data by handling missing values, normalizing numerical features, and encoding categorical variables.Engineer domain-specific features such as claim severity, historical risk indicators, and claim frequency to enhance predictive performance.Build and evaluate machine learning models including Decision Trees, Random Forest, Gradient Boosting, and XGBoost for risk classification and anomaly detection.Address class imbalance using SMOTE and class weighting, improving detection of high-risk cases.Optimize model performance through hyperparameter tuning and assess models using metrics like precision, recall, F1-score, and ROC-AUC.Deploy models into production to automate risk-based decision-making and monitor performance with periodic retraining.Develop dashboards and visualizations to communicate insights and recommendations to stakeholders.Provide actionable insights that improve process efficiency, reduce operational risk, and support data-driven decisions.Tech & Tools: Python, Pandas, NumPy, SQL, Scikit-learn, XGBoost, Random Forest, Decision Trees, Gradient Boosting, Feature Engineering, EDA, SMOTE, Dashboards & Reporting