Hi, my name is

• Developing AI solutions for financial data analysis and reporting in the finance sector
• Building multi-agent systems to automate and enhance financial decision-making processes
• Creating intelligent interfaces for executive-level financial insights and communication
• Integrating internal financial systems with AI capabilities to deliver real-time analytics

• Architected and deployed production-grade ML pipelines for request classification on GCP, achieving 47%

reduction in operational costs through automated workflow optimization and efficient resource utilization

• Designed and implemented a RAG-based chat application processing guest inquiries, achieving 50%

improvement in customer satisfaction and optimizing deployment through GCP services (Cloud Run)

• Implemented advanced LLM optimization techniques including fine-tuning and quantization, delivering 3X

faster inference times while maintaining model accuracy

• Engineered an intelligent email processing system using LangChain and custom LLMs, resulting in 60%

improvement in request processing accuracy

• Developed a Vision LLM pipeline for hotel room categorization with 99.9% uptime, achieving 40%

efficiency gain through automated feature extraction and classification

• Developed and deployed in-house solution for request classification, cutting costs by 47%.
• Optimized, fine-tuned, Quantized Language Models for Enhanced Performance, inference by up to 3X.
• Built an email parser API for guest request processing, automating data extraction from emails and reducing processing time by 60%.
• Created a RAG chat-app for quick and accurate resolution of guest inquiries, reducing response time by 50%.
• Employed vision LLM for automated hotel room and feature categorization, leading to a 40% increase in efficiency.

• Designed a facial expression recognition system using the FER2013 dataset, achieving an accuracy of 85%.
• Utilized SOTA computer vision models and evaluated their performance, achieving an average precision of 90%.
• Improved model performance by 20% through data augmentation, and regularization techniques etc.
• Incorporated the model into a desktop app, resulting in a 30% increase in user engagement and satisfaction.