Neha D.

About Neha D.

GenAI Engineer with 4.4+ years of experience building and deploying production-grade AI…

Experience

1. System Engineer

   Infosys

   Dec 2022 — Present · Pune, IN

   Project: Generative AI-Powered Chatbot for Field Engineers - British Telecom• Built a Retrieval-Augmented Generation (RAG) pipeline using LangChain and AWS Bedrock (LLaMA 2), leveragingPinecone as a vector database for semantic search and document retrieval.• Implemented document ingestion, embedding generation, and vector indexing workflows to enable context-awareand document-grounded response generation.• Used Amazon S3 for scalable document storage and retrieval.• Mitigated hallucinations through prompt engineering, retrieval optimization, and response validation mechanisms.• Used LangSmith for performance tracing, evaluation, and observability of LLM responses.• Impact: Automated enterprise document access, reduced manual effort, and improved field engineer productivity.Project: RCA Agent• Developed an AI-driven Root Cause Analysis (RCA) agent in Snowflake to investigate $500M+ inventory exposureacross supply chain and commercial datasets.• Designed RCA Trigger and Baseline layers using Snowflake materialized views to detect abnormal inventory riskacross SKU, distribution center, aging days, and exposure value metrics.• Leveraged Snowflake Cortex Analyst to automate hypothesis generation and evidence benchmarking against his-torical KPIs including forecast error, conversion rate, and lifecycle lag.• Transformed raw transactional inventory and commercial data into KPI-driven analytical signals to enable hypothesis-based RCA evaluation.• Implemented attribution modeling and automated narrative generation logic to identify key business drivers anddeliver actionable RCA insights to stakeholders