Databricks
About this role
A rare Databricks posting at 3 years, and it is squarely on the AI agent work everyone is hiring for. The job is building and deploying LLM powered agents end to end: the agentic logic in Python, a high performance React interface, middleware in Python or Node.js to orchestrate agent calls, and the observability and tracing to see what an agent run actually did. The posting names LangChain and LangGraph and asks for ReAct style reasoning loops specifically. The one requirement that will filter most applicants is not the years figure: it asks for proven experience designing, building and deploying at least one production deployed AI agent or LLM application. Side projects and demos will not clear that bar, shipped systems will. Bengaluru.
Who this is for
Required by the posting: 3+ years of experience in full stack development with strong expertise in Python and React.js. Proven experience designing, building and deploying at least one production deployed AI agent or LLM application. Deep expertise in LLM orchestration, prompt engineering and using tools and plugins to extend agent capabilities. Experience with agent frameworks like LangChain or LangGraph including implementing reasoning patterns such as the ReAct loop. Experience with observability, logging and tracing for AI agents. Experience building large scale, enterprise grade applications. Strong experience with API integration over REST or GraphQL. Experience working in Agile or Scrum environments.
Preferred or good to have: Node.js or Python for middleware development, understanding of Salesforce APIs including REST, Apex and RLM/CPQ, experience with event driven architectures such as webhooks, change data capture and async processing, knowledge of CI/CD pipelines and DevOps practices, and experience with the Databricks platform and Mosaic AI.
The real day to day: designing, developing and deploying AI agents for business critical workflows with a focus on LLM orchestration and tools; implementing agentic behaviours with LangChain or LangGraph including ReAct loops; building a high performance React interface for the agent; integrating APIs and services with backend teams and building tools for agents; orchestrating agent calls and data flow through Python or Node.js middleware; implementing observability and tracing for agent runs; building reusable UI components; and contributing to architectural discussions.
Location: Bengaluru, India, in Databricks India Private Limited.
Honest fit guidance: the years bar and the real bar are different here, and it is worth being clear about that. Three years is genuinely low for Databricks, where most of today's other India requisitions start at 10 or more. But "at least 1 production-deployed AI agent or LLM application" is a hard, checkable requirement, and the Salesforce API preference suggests this agent work plugs into internal business systems rather than the core data platform. If you have shipped an agent into production at any scale, your years matter less than that fact.
Preferred or good to have: Node.js or Python for middleware development, understanding of Salesforce APIs including REST, Apex and RLM/CPQ, experience with event driven architectures such as webhooks, change data capture and async processing, knowledge of CI/CD pipelines and DevOps practices, and experience with the Databricks platform and Mosaic AI.
The real day to day: designing, developing and deploying AI agents for business critical workflows with a focus on LLM orchestration and tools; implementing agentic behaviours with LangChain or LangGraph including ReAct loops; building a high performance React interface for the agent; integrating APIs and services with backend teams and building tools for agents; orchestrating agent calls and data flow through Python or Node.js middleware; implementing observability and tracing for agent runs; building reusable UI components; and contributing to architectural discussions.
Location: Bengaluru, India, in Databricks India Private Limited.
Honest fit guidance: the years bar and the real bar are different here, and it is worth being clear about that. Three years is genuinely low for Databricks, where most of today's other India requisitions start at 10 or more. But "at least 1 production-deployed AI agent or LLM application" is a hard, checkable requirement, and the Salesforce API preference suggests this agent work plugs into internal business systems rather than the core data platform. If you have shipped an agent into production at any scale, your years matter less than that fact.
Apply on company site
Opens databricks.com, the employer's own application page. Applying is always free.