Elan is a platform for building, deploying and scaling AI agents. A developer writes an agent, and the platform provides everything else needed to run it as a product. You would be one of the first two engineers on the team.
As Platform Engineer you own the agent runtime and every shared service an agent depends on. Their design, quality and uptime are yours.
What you will own
Agent runtime and SDK. The libraries every agent on the platform is built on. They start in Python and extend to TypeScript on Node.js and to Go, with support for agent frameworks such as Google ADK and LangGraph.
Identity and real-time delivery. Token verification with JWT and JWKS, session identity, and real-time streaming between an agent and its users.
Usage ledger. Usage and balance records that have to be exactly right, with reconciliation.
Observability. Run history and an audit trail for every agent run, kept separate per tenant.
Policy and approvals. A service that decides what an agent is allowed to do and routes the rest to a human for approval.
Web console. A React and TypeScript application where customers see usage, runs, approvals and policy.
Data layer. PostgreSQL schemas, versioned migrations, and event-driven messaging on Pub/Sub with per-tenant isolation.
How you will work
You report to the head of product and technology.
You work beside a DevOps Engineer who owns the deploy path and the infrastructure. The two of you alternate on-call weeks.
You talk to the developers who build on the platform. Their problems set your priorities.
You are based in Noida with a hybrid working arrangement. Compensation is 40 to 90 LPA.
We hire for four things: ownership, curiosity, developer obsession and delivering results.
Minimum requirements
8 to 15 years of backend engineering, with at least 3 building and running Python services in production.
Working proficiency in TypeScript on Node.js or in Go, enough to build and maintain an SDK in that language.
Has built or operated LLM agents in production with at least one agent framework, such as Google ADK or LangGraph.
Has designed an SDK, library or API that other developers build on, and versioned it without breaking them.
PostgreSQL in depth: schema design, migrations, query tuning and connection pooling.
Event-driven systems on Pub/Sub, SNS/SQS or Kafka, including idempotency and at-least-once delivery.
Authentication and authorization: OAuth 2.0 and OIDC, JWT and JWKS verification, role and tenant scoping.
Has shipped a multi-tenant product and can explain how tenant isolation was enforced in data and messaging.
Working React and TypeScript, enough to ship console features without a frontend partner.
Extensive production experience with Kubernetes: building, deploying, scaling and debugging services on it, including workload identity, resource limits, autoscaling and namespace isolation.
Writes tests as part of the change, and has owned a service through production incidents.
Good to have
Ledger, billing or payments systems.
WebSockets or other streaming protocols at scale.
Sandboxed container runtimes such as gVisor, and running untrusted workloads on Kubernetes.
Open-source maintenance, or developer-facing writing such as documentation and examples.
Experience as an early engineer at a startup.
GCP.
