Looking for a job

DATTATRAY DUBILE

@dattatraydubile

Agentic AI Engineer at Capgemini building multi-agent platforms and RAG workflows across AWS and Azure.

At Capgemini, I develop agentic AI and LLM solutions, including a multi-agent platform on AWS Bedrock using LangGraph and LangChain. It supports RAG and non-RAG workflows for code generation, document creation, and knowledge retrieval.

For Capgemini (Optimus AI, IDEA), I established a Supervisor–Sub-Agent architecture with A2A communication across AWS Bedrock, Azure OpenAI, and Vertex AI. I also developed a code generation system supporting six languages.

For Capgemini (Allianz), I built batch and real-time ETL/ELT pipelines across AWS and Azure, including an ingestion framework that improved performance by 30–40%. I also automated workflows that reduced manual intervention by 90%.

Earlier, for Capgemini (SDH), I delivered a domain-driven chatbot that achieved 95% accuracy and created a document ingestion pipeline combining OCR, vector embeddings, and a Neo4j knowledge graph.

Experience

Work history, roles, and key accomplishments

Current

Agentic AI Engineer

Current

Capgemini

Jan 2025 - Present (1 year 9 months)

Established Supervisor–Sub-Agent architecture with A2A communication, enabling cross-platform interaction across AWS Bedrock, Azure OpenAI, and Vertex AI via FastAPI services. Led development of multi-agent AI platform on AWS Bedrock using LangGraph and LangChain, enabling RAG and non-RAG workflows for code generation, BRD/HLD creation, and knowledge retrieval.

A2A Protocol LangGraph Lang Chain AWS Azure OpenAI Vertex AI fastAPI RAG Agentic AI Multi Agent Systems Stateful Workflows S3 Azure Blob Storage Code Generation Error Handling Retry Mechanisms Self Healing Systems

Current

Agentic AI Engineer / Data Engineer

Current

Capgemini

May 2024 - Present (2 years 5 months)

Built scalable ETL/ELT pipelines using the Medallion Architecture (Bronze, Silver, Gold) across AWS and Azure, supporting both batch and real-time streaming ingestion. Automated workflows using AWS Step Functions and JWT-secured APIs, reducing manual intervention by 90%.

ETL ELT Medallion Architecture AWS Azure AWS Glue Azure Databricks AWS Step Functions JWT REST APIs RAG Azure AI Foundry Databricks Unity Catalog CoPilot Studio SQL Generation LLM Natural Language Processing

Gen AI Engineer

Capgemini

Apr 2023 - May 2024 (1 year 1 month)

Delivered domain-driven chatbot using Azure OpenAI with multi-agent orchestration for intelligent query classification, domain validation, and context-aware response generation, achieving 95% accuracy. Created scalable document ingestion pipeline with OCR, hash-based deduplication, vector embeddings, and Neo4j-based knowledge graph for hybrid retrieval.