Sitemap
Medium Logo

==> A Post-Graduate in Management and B-Tech in Computer Science with 12+ years of hands-on experience in end-to-end project life cycle i.e., requirement gathering, POC execution, value projection, proposal submission and project delivery. Currently having 6 years of experience in leading client delivery, engagement and team management in Artificial Intelligence work space. ==> Demonstrated value stories and implementation benefits through AI use cases in the areas of Financial Crime (KYC & Transaction Monitoring), Marketing Analytics & Credit Risk (Credit Cards), Demand Forecasting (Retail), Process Automation (Insurance). ==> Experienced in building, scheduling, and monitoring MLOps orchestration pipelines using Azure Data Factory (ADF). Primarily utilized PySpark within Azure Databricks notebooks to develop ETL pipelines for incremental data ingestion into Azure Data Lake. Additionally, leveraged Python within Azure Databricks notebook to train various Machine Learning and Deep Learning models. Implemented CI/CD pipelines for seamless deployment and operational monitoring using Azure DevOps. Further used MLflow in Azure Machine Learning to monitor the model performance. ==> Experienced in developing Retrieval-Augmented Generation (RAG) based Multi-Agent AI workflows using orchestration framework like LangChain. Used LlamaIndex to preprocess (chunking, indexing and embedding) various forms of unstructured data and then storing it into VectorDBs like ChromaDB, FAISS, Pinecone for accurate retrieval of information. Leveraged various open source LLMs like Llama 3-8B, Mistral-7B as well as commercial LLMs like GPT-4, GPT-4o to generate responses across diverse range of language tasks.

Debasish Dutta

Debasish Dutta

Data Science + Generative AI + Azure MLOps || 12+ Years || Financial Services (Retail Banking, Wealth & Asset Management) + Retail (FMCG, CPG)