Hire K K. - RAG Engineer

* Architected and deployed an enterprise-grade Hotel Recommendation AI Chatbot using RAG pipelines (LangChain * FAISS * Gemini), processing real-time guest queries across 5,000 * hotel properties — reduced query resolution time by 45%. Built an Intelligent Document Intelligence System for internal HR and operations teams using LLM-powered extraction, classifying and processing 10,000 * unstructured documents monthly with 92% accuracy. Designed a Knowledge Assistant Bot for ACCOR's internal IT helpdesk using LlamaIndex * OpenAI GPT-4, cutting average ticket resolution time by 35% and reducing L1 support escalations. Developed end-to-end MLOps pipelines using MLflow, Kubeflow, and Docker/Kubernetes on Azure ML — enabling automated retraining, A/B testing, and model drift monitoring in production. Led technical design reviews, sprint planning, and mentored 3 junior engineers on GenAI best practices and prompt engineering.

Designed and deployed NLP-driven Customer Sentiment Analysis pipeline using BERT embeddings + Scikit-learn, ingesting 500K+ product reviews monthly to drive marketing strategy decisions. Built a real-time MLOps workflow using Apache Airflow + MLflow for automated model training, versioning, and deployment — reducing model release cycles from 2 weeks to 3 days. Developed predictive analytics models for sales forecasting and customer churn prediction using XGBoost and LightGBM, achieving 87% prediction accuracy. Created interactive Tableau dashboards for C-suite stakeholders visualizing KPIs, campaign ROI, and model performance metrics.

* Investigated seller fraud, counterfeit product listings, and policy violations using data-driven risk scoring models — contributing to a 20% reduction in fraudulent seller activity. * Analyzed large transaction datasets using SQL and Python to identify anomalous patterns and escalate high-risk cases. * Collaborated with ML teams to validate and provide feedback on risk detection models, improving false-positive rates. * Developed automation scripts in Python to streamline repetitive investigation workflows, improving team productivity by 30%.