Principal AI Engineer - Hybrid in Bangalore
About this role
To address the need for a cohesive infrastructure that unites data, knowledge graphs, and AI agents, Smartsheet requires a Principal AI Engineer to establish core technical standards at their Bangalore hub. The engineering team focuses on embedding artificial intelligence into the intelligent work management platform to automate manual tasks and scale enterprise insights. In this capacity, the individual creates the target-state architecture for the AI platform, designs the agent harness and sub-agent framework, and builds reference agent frameworks that internal teams can extend. Ideal candidates possess a strong background in architecting production agentic systems, excel at rapid prototyping to de-risk complex technical decisions, and can drive cross-team alignment through design quality and working software rather than positional authority.
What you will do
- Own the target-state architecture for the AI platform by defining standards that bind data, agents, knowledge graphs, and context graphs into a single system.
- Design the agent harness and sub-agent framework encompassing orchestration, lifecycle management, hand-off contracts, evaluations, and capability-tier classifications.
- Develop schemas, ontologies, ingestion pipelines, and retrieval patterns to ensure agent responses remain grounded and reliable.
- Build and maintain a reference agent framework that development teams can fork or extend without rebuilding orchestration or tracing from scratch.
- Author and publish a shared skills and tools library equipped with versioned interface contracts for broader organizational adoption.
- Develop fast prototypes to de-risk architectural decisions regarding retrieval strategies and partner capabilities before committing to roadmaps.
- Establish deployment, security, and governance patterns to standardise the path from prototype to production across all platforms.
- Act as the primary technical lead for strategic technology partners including Databricks, AWS, and Anthropic by running technical syncs and evaluations.
What you need
- Deep, hands-on experience architecting production agentic systems on the Databricks Platform, AWS Bedrock, AgentCore, and Model Context Protocols.
- Demonstrated background in agent harness and sub-agent orchestration design, tracing, observability, evaluations, governance, and cost instrumentation at scale.
- Strong experience working with knowledge and context graph based systems, covering ontology design, ingestion pipelines, and retrieval evaluation.
- A proven track record of prototyping rapidly, shipping working software, and driving adoption through utility rather than mandate.
- Experience designing reusable, versioned software interfaces and libraries with clear promotion criteria from prototype to production.
- Familiarity with modern LLM and agent tooling such as Strands SDK, Databricks, and Bedrock.
- Ability to serve as a technical DRI in strategic partner relationships and translate external roadmaps into internal platform decisions.
- Legally eligible to work in India on an ongoing basis.
Skills and tech stack
Questions candidates ask
What is the location policy for this role?
This position is a hybrid role based in Bangalore, India at the company's Global Capability Center.
What technologies and platforms will I work with?
The tech stack includes Databricks, AWS Bedrock and AgentCore, Anthropic Claude, MLflow, and Model Context Protocols (MCP).
What will I actually be working on day-to-day?
You will own the target-state architecture for the AI platform, design agent harnesses and sub-agent frameworks, build reference agent frameworks, and partner with technology vendors.