Data Scientist Generative Ai

Rohan Varma Bandari

+1-940-***-**** *****************@*****.*** http://www.linkedin.com/in/rohanbandari Summary

Data Scientist & Generative AI Engineer with 3+ years of experience delivering enterprise-scale AI solutions across Finance, Healthcare, and Retail. Expert in Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), prompt engineering, and full-stack ML/AI pipeline development. Adept in designing explainable, scalable, and cloud-native AI architectures using AWS, Azure, and GCP.

Technical Skills

• Programming Languages: Python, SQL

• Development Tools: Jupyter Notebook, JupyterLab, Amazon SageMaker, Google Colab, VS Code, PyCharm

• Data Analysis & Visualization: NumPy, Pandas, Matplotlib, Seaborn, SciPy, Power BI, Tableau, Plotly, Excel

• Deep Learning: CNNs, RNNs, LSTM, VAEs, GANs, Attention Mechanisms, Self-Attention, PyTorch, TensorFlow

• Generative AI: OpenAI LLMs, Hugging Face, LangChain, RAG, AutoGen, CrewAI, LangGraph

• Vector Databases: Chroma, Faiss, Milvus, Pinecone

• Prompt Engineering: Chain of Thought (CoT), Few-shot Prompting

• MLOps & Pipelines: MLflow, Kedro, Apache Airflow, Docker, Kubernetes, FastAPI

• Cloud Platforms: AWS (S3, SageMaker), Azure (OpenAI, AKS), Google Cloud Platform (Vertex AI)

• Database Systems & Version Control: MongoDB, NoSQL, Snowflake, Git, Bitbucket

• Explainability & Monitoring: SHAP, LIME, AUC-ROC

• Machine Learning & Forecasting: Scikit-learn, XGBoost, LightGBM, Random Forest, Logistic Regression, Autoencoders, CatBoost, ARIMA, SARIMA, SARIMAX, Prophet, TFT

• Natural Language Processing: NLTK, spaCy, Gensim, Transformers, BERT, GPT, NER, Sentiment Analysis, OpenAI Embeddings, Chatbots, Information Extraction

Professional Experience

Wells Fargo May 2024 - Present

Data Scientist / Generative AI Engineer USA

• Developed GenAI-powered risk analysis & compliance system using LangChain, RAG, and CrewAI to streamline regulatory document processing.

• Integrated Whisper and Azure Form Recognizer for multimodal entity extraction from audio and documents, enhancing data coverage.

• Designed multi-agent pipelines for fraud detection and regulatory insights in unstructured financial data.

• Fine-tuned GPT-4o, Llama 3 70B with LoRA, QLoRA for cost-efficient financial modeling.

• Collaborated with compliance analysts to automate SQL-based reporting and developed Tableau dashboards for real-time audit insights.

• Reduced document review time by 75% and increased fraud risk prediction accuracy by 50%. Omega Healthcare April 2022- July 2023

Data Scientist (ML) India

• Built fraud detection models for insurance claims using XGBoost, LightGBM, and Autoencoders, enabling early flagging of suspicious submissions across large-scale structured and unstructured data.

• Integrated SHAP-based explainability into scoring pipelines, helping claims analysts interpret and act on model decisions.

• Deployed end-to-end ML pipelines on AWS SageMaker with real-time scoring and periodic retraining based on fraud trend data.

• Designed custom anomaly detection modules to trigger audit flags pre-payout, reducing false positives and claim review delays.

• Built HIPAA-compliant analytics pipelines and dashboards that monitored for claims-based patient outcome tracking and operational efficiency, improving clinical resource allocation and reducing wait times by 22%.

• Achieved a 60% drop in fraudulent claims and a 35% boost in processing efficiency. Aditya Birla Retail June 2021 - Mar 2022

Associate Data Scientist India

• Built NLP pipelines for sentiment analysis, aspect-based sentiment analysis (ABSA), and extraction of product-level insights from customer reviews.

• Developed personalized product recommendation engines using CNNs and BERT embeddings to boost engagement and cross-category discovery.

• Created deep learning–based demand forecasting models using LSTM and Temporal Fusion Transformers to optimize store-level inventory and purchasing decisions.

• Analyzed over 10M+ customer transactions and loyalty records with pandas and SQL to identify purchase patterns, driving a targeted upsell campaign that increased average basket size by 18%.

• Designed and deployed filters to detect and reduce fake reviews, cutting fraudulent review influence by 50% across key product lines.

• Improved inventory turnover and supply planning efficiency by 35% through accurate time-series predictions and retail seasonality modeling.

Key Projects

• Financial Compliance GenAI System (Wells Fargo): Built LLM-powered RAG pipeline with LangChain & CrewAI, reducing audit workload by 75%.

• Insurance Fraud Detection (Omega Healthcare): Deployed SHAP-based XGBoost models on SageMaker for real-time claims fraud scoring.

• Retail Demand Forecasting (Aditya Birla): Designed LSTM/TFT forecasting system boosting inventory accuracy by 35%.

Certifications

• Microsoft Certified: Azure AI Engineer Associate

• Oracle: Database Programming with SQL

Publications

Question Answering System using NLP and ALBERT

• Published a paper on Building a QA system that leverages advanced NLP techniques and ALBERT model optimizations (Sentence Order Prediction, Parameter Sharing) to deliver accurate responses to natural language queries. Focused on linguistic analysis, information retrieval, and real-world applications. Highlights skills in model fine-tuning, dataset preparation, and Python-based implementation relevant to GenAI and intelligent data systems. Education

University of North Texas

Master of Science, Computer Science

Vidya Jyothi Institute of Technology

Bachelor of Technology, Computer Science and Engineering

Contact this candidate