Devadharshani A

At WNS Analytics, I build AI solutions for insurance claims, including a LightGBM salvage-cost prediction module that contributed to 81% overall total-loss prediction accuracy and significant claim-cost savings.

I also led engineering-report analysis to identify cost inefficiencies, finding that 69% of non-payout claim expenses were driven by engineering reports. I designed an AI-based OCR framework to extract cause-of-damage and severity information so businesses can act proactively.

Previously at NeoStats Analytics, I led delivery of production AI products including a judicial modernization platform using Azure OpenAI, reducing judgment drafting from 3–4 weeks to 2–3 days. As Technical Lead for an insurance claims processor, I automated validation, entity extraction, identity verification, and claim decisions, reducing processing time by 87%.

I work across Generative AI, LLMs, RAG, NLP, Azure, Databricks, and Power BI, and have led three production projects from initiation through deployment. I bring a data-science foundation, hands-on engineering experience, and a focus on measurable operational impact.