Darshan Padaliya

CRM BOT: Developed a bot integrated within ours CRM platform to optimize user task efficiency and productivity. Implemented intent classification and slot filling for seamless understanding and execution of user commands. Supported both text and audio inputs for enhanced user accessibility and convenience. Defined multiple intents such as scheduling meetings, updating tasks, open contacts etc. to cater to user needs. Integrated open source LLM for generating responses if user’s command doesn’t fall into any of the intents. Added RAG pipeline to enhance the chatbot's knowledge base functionality. ENTITY RESOLUTION: Created a robust record linkage system. Vectorized records and formed clusters of similar records based on cosine similarity matrix. Built a scalable system that can easily handle large datasets of up to billions of records. INVENTORY FORCASTING: Development of an inventory forecasting system for retail stores, utilizing Databricks for lakehouse architecture. Focused on data cleansing and feature engineering including fetching, cleaning, transforming more than million records per batch with incremental load to optimize model accuracy. Wrote data test cases in CI/CD pipeline to ensure model performance and accuracy. Touched millions of customers with the forecasting system, improving inventory management and reducing stockouts.