Rahul Pradhan Email and Phone Number

• Built and delivered end-to-end media insights and highlights/tags pipeline for creating the media insights from the data generated from different emotion models for multi-modality insights use cases. • Developed and implemented an automated report generation module leveraging the flexibility and power of LLM for both quantitative and qualitative aspects for the market research platform decode, hosted as a packaged solution in CodeArtifact.• Maintained Speech-to-text pipeline where using different vendors(Google, AzureSTT, AWSSpeech, Transkriptor, Speechmatics and deepgram) generated text from speech. • Collaborated across teams as the sole resource responsible for MLOPs. • Showcased expertise in optimizing many existing jobs and services such as introducing parallel computing from pandas to dask dataframe and multi-threading. • Managed and maintained the CI/CD pipeline for various tasks, including step function, AWS Lambda and Batch Jobs.• Built and delivered a text analytics engine for Decode, with capabilities such as text summarization, theme detection, and sentiment analysis by using NLP and fine-tunning the LLM for market research purposes and leveraging Docker and AWS batch for deployment.• Built a custom Retrieval Augmented Generation (RAG) system for working with consumer insightsdata using LangChain and open-source large language models (LLMs). • Collaborated on the integration of an AI Copilot for both quantitative and qualitative analysis on base of AWS Lex, optimizing user interactions and augmenting the product’s analytical capabilities using LLMS and LangChain.