Agentic AI

As an Agentic AI Developer on our core Agentic AI Engineering team, you will be tasked with building modular, prompt-based GenAI agents that handle tasks such as spec parsing, validation, DDL generation, and metadata enrichment. Your responsibilities will include implementing multi-step intelligent agents that seamlessly integrate into the data pipeline to automate manual, context-heavy tasks using LLMs and structured logic. Key Responsibilities: - Build and maintain multi-task agents utilizing frameworks like LangChain or Bedrock Agents - Develop sub-services for spec parsing (Excel/CSV to JSON), validation rule generation (from metadata), and DDL generators for Redshift, BigQuery, and Snowflake - Integrate LLM outputs with structured data systems such as dbt, Postgres, and S3 - Implement agent lifecycle management including versioning, logging, and pre/post hooks - Collaborate with data engineers to ensure the usability of agent outputs in ingestion pipelines - Contribute to prompt design, evaluation scripts, and testing harnesses for LLMs - Participate in the continuous refinement of prompt pipelines such as Chain-of-Thought, ReAct, and RAG Tech Stack: - LLM Platforms: OpenAI, Bedrock, Anthropic - Prompting: LangChain, LlamaIndex, PromptLayer - Backend: Python, FastAPI, SQLAlchemy - Data: dbt, PySpark, Redshift, Snowflake - Infra: GitHub, Docker, S3, Terraform - Validation: Great Expectations, dbt Tests (generated dynamically) Qualifications Required: - 2-4 years of experience in Python backend or data engineering roles - Experience with GenAI prompt integration, even in hobby or PoC projects - Strong understanding of JSON schema, YAML configs, and spec-driven development - Knowledge of data validation, test automation, or SQL generation from metadata - Familiarity with agents, chains, LangChain tools, or Bedrock workflows Please note that this is a 3-6 months contract position based in Pune. Work will be carried out in the office (WFO).,