Hire Sanjana D. - Generative AI Engineer

* Built GenAI solution to auto-generate compliance reports from regulatory filings. * Developed GPT-based assistant for internal audit and financial reconciliation. * Created document classification LLM for fraud detection workflows. * Integrated OpenAI with banking CRMs to reduce manual data entry. * Fine-tuned LLaMA model on financial transaction datasets for anomaly detection. * Designed GenAI chatbot to answer regulatory and policy questions for employees. * Used embeddings for contract search and risk clause extraction. * Developed secure RAG architecture with bank-grade encryption. * Applied LLMs to generate synthetic financial datasets for model testing. * Deployed voice-to-text summarization for wealth management call centers. * Built AI co-pilot to guide KYC/AML analysts through due diligence workflows. * Used LangChain to combine SQL database access with natural language queries. * Created LLM-based tools for Basel III stress testing document analysis. * Enhanced fraud team workflows with GPT-powered summarization of transaction trails. * Built multi-agent system for investment portfolio generation using financial LLMs. * Integrated AI systems with Microsoft Azure’s compliance and auditing frameworks. * Helped reduce onboarding time by 30% using GenAI form completion tools. * Implemented prompt engineering playbooks for banking-specific risk terms. * Developed hybrid models combining traditional ML and transformer layers. * Built custom embeddings for retail and commercial banking product catalogs. * Automated SWIFT message interpretation using fine-tuned transformer models. * Achieved 20% faster loan processing using AI-driven document parsing. * Used vector search for regulatory clause comparison across regions. * Deployed scalable AI inference pipelines in cloud-native architecture. * Trained bank employees on using AI copilots in day-to-day operations.

* Built AI assistant to summarize claims adjuster notes into structured claims data. * Developed OCR + GPT pipeline to digitize and interpret handwritten claim forms. * Integrated LLM into underwriting workflows for faster policy issuance. * Created a chatbot to guide agents through product recommendations and FAQ. * Trained transformer models on large P&C (Property & Casualty) datasets. * Used GenAI to automate first notice of loss (FNOL) intake process. * Designed document clustering tools for group policy management. * Built secure LLM interface for internal knowledge base of insurance policies. * Created GenAI-powered chatbot to assist with fraud triage. * Integrated AI insights with Salesforce Insurance Cloud. * Created synthetic insurance claim datasets using diffusion + LLM for testing. * Applied RAG to summarize multi-document policy bundles. * Developed custom summarization models for reinsurance reports. * Delivered 40% improvement in quote generation time using prompt-based models. * Used LangChain agents for structured Q&A from complex policy PDFs. * Built LLM validator system to flag hallucinated responses. * Enhanced customer onboarding journey with personalized AI content generation. * Applied rule-based post-processing to align LLM output with underwriting rules. * Collaborated with actuaries to ensure AI compliance with regulatory models. * Delivered multilingual support via LLMs for claims in Spanish and French. * Embedded AI models into mobile app workflows for field agents. * Leveraged FAISS for similarity search on historical claim precedents. * Built explainable scoring models for insurance premium calculation. * Automated policy renewal notices using fine-tuned GPT model. * Designed human-in-the-loop feedback systems to improve LLM response quality.

* Built a clinical NLP pipeline to extract structured data from unstructured doctor notes. * Deployed HIPAA-compliant ML models for predicting patient readmission risk. * Integrated machine learning modules with Cerner EHR via FHIR API. * Used BERT and BioGPT models for summarizing long-form medical notes. * Developed an AI-based system to flag adverse drug reaction cases from EMR notes. * Created a FastAPI app for internal users to query patient trends using natural language. * Implemented time-series models for ICU patient vitals forecasting. * Worked with the radiology team to classify X-rays using ResNet and ViT models. * Built automated dashboards in Streamlit to visualize patient data predictions. * Applied SHAP and LIME to ensure model explainability for clinicians. * Trained custom NER model to detect symptoms, medications, and conditions. * Created multi-class classification models for disease progression stages. * Reduced model training time by 40% using mixed-precision training. * Optimized hyperparameters using Ray Tune and Optuna. * Implemented CI/CD pipeline for deploying updated models weekly. * Used Azure ML for versioning, deployment, and monitoring of health ML models. * Created a model registry for managing ML lifecycle across departments. * Built a summarizer tool to convert doctors’ voice notes to brief structured summaries. * Performed drift detection and retraining scheduling on incoming EHR data. * Used unsupervised learning to detect unusual trends in outpatient care data. * Delivered workshops for clinical staff on interpreting AI predictions. * Designed and deployed a chatbot assistant for clinicians to ask clinical protocol questions. * Integrated privacy-preserving methods like differential privacy for sensitive datasets. * Created a data lake with compliant access control for ML training purposes. * Worked cross-functionally with data engineers, compliance officers, and clinicians.