Arpana Nyaupane
Data Scientist, Ai @MetLife
Signup · Get unlimited contacts
WORK HISTORY
Data Scientist, Ai @MetLife
CA
Architected a Retrieval-Augmented Generation (RAG) system using LangChain and LlamaIndex to automate the interpretation of complex Canadian insurance policy documents, reducing manual inquiry volume by 40% and improving response accuracy to 96%. • Fine-tuned open-source Llama-3 models using LoRA and PEFT techniques on anonymized medical claim data to automate the initial triage process, resulting in a 35% acceleration in claims processing times. • Developed a gradient-boosted ensemble model using XGBoost and PySpark to predict CLV across diverse product lines, enabling personalized marketing strategies that increased cross-sell conversion rates by 22%. • Spearheaded the migration of legacy ML models to a containerized Kubernetes environment on Azure ML, utilizing Terraform for Infrastructure-as-Code (IaC) to reduce deployment lead time from weeks to hours. • Engineered a deep learning-based churn prediction engine using PyTorch, identifying high-risk policyholders with a 0.89 ROC-AUC, allowing the retention team to mitigate $2M in potential annual premium loss. • Designed a robust feature store in Snowflake using SQL and Python, centralizing 500+ engineered variables for real-time model scoring, which enhanced model training efficiency by 50%. • Conducted large-scale A/B Testing on the digital member portal’s recommendation engine, utilizing Bayesian inference to validate a 12% lift in user engagement with supplemental health products.
EDUCATION
Eastern University
Master of Science - MS, Data science
Southern Alberta Institute of Technology (SAIT)
Bachelor's degree, Bachelor degree in geographic information system
ABOUT ARPANA NYAUPANE
Data Analyst with 4+ years of experience Data Scientist, delivering end-to-end machine learning and Generative AI solutions within the financial and insurance sectors. I specialize in building scalable AI systems, including Retrieval-Augmented Generation (RAG) pipelines, NLP models, and predictive analytics solutions that drive measurable business impact. Currently, I focus on deploying production-grade ML systems using modern MLOps practices with Docker, Kubernetes, and Terraform. I enjoy bridging the gap between advanced statistical modeling and real-world deployment at scale. I’m passionate about solving complex data problems and building intelligent systems that create real business value.
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.