AUMKAR BHAMANGOL

Developed and deployed end-to-end data ingestion pipeline from MQTT -> Kafka -> Feature Engineering -> Postgres Database. Worked with GPT4o, gpt3.5, AzureOpenAi, Phi-2, Multiagent Framework such as CrewAi, Autogen, LlamaIndex, Langchain, Langgraph. Engineered features from timeseries data for monitoring the workforce's proximity on shopfloor increase productivity by 60%. Build custom NLP-to-SQL solution using LlamaIndex & Autogen for communicating with enterprise tabular data in natural language using OpenAI with 84% accuracy. Developed custom RAG for interacting with Machine Operating Manuals PDF (Error code) helping managers to resolve problems 90% faster.

Developed Apori Algorithm for Basket Analysis for cross-selling based on items in Basket increasing sales by 23%. Successfully predicting optimal product stocking strategy based on seasonality and trends reducing CAPEX investment by 40%. Implemented Dynamic Discounting on products to increase the churn rate in stores by 15%.