Training and Evaluation for Responses from Large Language Models.Core Understanding of response and providing significant improvement for Model performance compared with SOTA models.Improved LLM performance by 25% using domain-specific prompt engineering for improved NLP alignment and context-awareness.Collaborated with cross-functional teams to design solutions for complex language tasks, showcasing adaptability and teamwork in a dynamic AI environment.

Developed and implemented Marketing Mix Modeling (MMM) using Python libraries such as Stats models and Scikit-learn to analyze the impact of different marketing strategies on hotel bookings and revenue, optimizing marketing spend.Implemented machine learning algorithms for collaborative filtering and matrix factorization to personalize hotel recommendations based on user preferences and booking history.Utilized natural language processing techniques like sentiment analysis and topic modeling to extract insights from customer reviews and feedback, enhancing the recommendation system's understanding of guest preferences and improving service quality.Explored deep learning architectures such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs) to analyze unstructured data like images and text, improving the accuracy of hotel image recognition and guest preference identification.Conducted research on transfer learning approaches to leverage pre-trained language models like GPT-3 and BERT for fine-tuning on domain-specific datasets, enabling the recommendation system to adapt to evolving guest trends and preferences.Employ techniques such as word embeddings and sequence-to-sequence models to generate personalized hotel descriptions or responses, enhancing the guest experience and engagement with the booking system.Integrated conversational AI algorithms like Rasa into the system architecture to develop a chatbot capable of engaging in natural language conversations with guests, providing real-time assistance, booking support, and tailored hotel recommendations.Utilizing Multi-Touch Attribution (MTA) models to evaluate the contribution of various marketing touchpoints across the guest booking journey using Python libraries. Analyze data to improve marketing strategies and enhance customer acquisition and retention.

Utilized Statistical techniques to solve complex problems and improve efficiency. Implemented Computer Vision solutions for Image processing, masking and Object Detection using Faster RCNN, Yolo and likewise pre-trained models, Image Classification/Segmentation. For a Sports Media based client where I was part of the Max-Auto-Optimization team. My primary responsibilities included constructing precise, scalable, and resilient prediction models using extensive datasets. These models were instrumental in real-time bidding and recommendation systems. Participated and contributed to sprint planning, estimation, and code reviews, advocating innovative solutions and long-term technical investments that will drive innovation. Developed and maintained Directed Acyclic Graphs (DAGs) in Apache Airflow to orchestrate complex workflows for data collection, preprocessing, model training, evaluation, and deployment. Collaborated with studying comparative performance between classical machine learning model and quantum computing models such as VQR and VQLSAutomation Engineer (Python)Spearheaded the automation of manual testing processes through the utilization of Python Pytest, thereby enhancing efficiency and accuracy in test execution.Design, develop, and maintain robust and scalable automated test frameworks to validate software functionality, ensuring a comprehensive test coverage.Monitor and analyze test results, generating comprehensive reports and metrics to aid in decision-making and process optimization.Applied deep learning and machine learning architectures, including transformers, large and small dedicated language models to solve business problems and speed up innovation and PyMC3 and Stan for Bayesian modeling and Hypothesis testing including A/B testing and other statistical methods.