Senior AI Engineer at Hyatt Hotels Corporation

Built and deployed PyTorch/TensorFlow models for guest personalization and operational optimization. Owned MLOps pipelines using Docker, Kubernetes, AWS SageMaker, GCP Vertex AI, and Airflow for automated training, deployment, and monitoring. Designed SQL/Airflow/Spark data pipelines for real-time analytics and AI-driven guest experience improvements. Integrated AI services into hospitality platforms via RESTful APIs. Implemented model drift detection and automated retraining workflows, and used MLflow/model registries for reproducible experiments and robust model lifecycle management. Reduced average response times by 35% and improved production uptime by optimizing inference performance and reliability; reduced incident resolution time by 40%. Developed LLM chatbots and RAG systems to improve engagement, reduce response/information retrieval time, and enhanced booking/retail recommendations to increase upsell conversion by 25%. Improved operational costs by 20% via workflow automatio

Senior Machine Learning Engineer at Urbint

Developed an end-to-end surveillance and risk analysis system integrating 200+ CCTV cameras on Google Cloud for real-time processing. Fine-tuned LLMs to create a chatbot-style recommendation system. Built document parsing using CV+NLP with Google Cloud serverless components (Cloud Functions, Pub/Sub, Cloud Run) and Vision/Natural Language APIs, plus custom ML deployment with AI Platform (Unified), model registry, and GKE to support horizontal scalability. Authored and executed testing (Cypress, Pytest) with up to 85% coverage. Architected a secure, privacy-first, multi-tenant GCP solution from conceptual design to component design, and mentored team members while collaborating across DevOps and development activities.

Machine Learning Engineer at Cisco

Built an NLP-powered product recommendation and classification pipeline using NER/NLU/NLG. Implemented keywords extraction and unsupervised LDA topic modeling for classifying service request cases. Classified Cisco products into categories to improve sales effectiveness. Clustered product features from text and summary fields using NLP methods and K-means clustering.