AI Engineer

Senior AI Engineer

Agentic Systems · Full-Stack (Go + TypeScript / Next.js)

Location: Mumbai / Bengaluru (Hybrid)

Compensation: ₹50–70 LPA + meaningful equity

(Compensation can be discussed for the right candidate)

Tech Stack: Go (orchestration engine) · TypeScript / Next.js (product) · AWS · In-house agent runtime (no LangChain)

What You'll Build

We're building an AI operator for real businesses—software where AI agents run Sales, Marketing, Operations, Finance, HR, and Legal workflows end-to-end, rather than simply suggesting next steps.

You'll own key systems that make this reliable and safe at scale: the agent orchestration engine, the multi-tenant platform, or the governance layer covering approvals, auditability, and guardrails.

The challenge isn't calling an LLM—it's ensuring agents behave predictably across thousands of tenants, under partial failure scenarios, without taking irreversible actions.

We build our agent runtime from scratch—no LangChain, LangGraph, or off-the-shelf frameworks. We directly integrate with model APIs and own every layer:

• Agent loop (reason → act → observe)
• Tool-call dispatch and schemas
• Context-window assembly
• Token budgeting
• Structured and constrained outputs
• Evaluation and regression frameworks

The orchestration engine is built in Go and runs agents as durable, checkpointed workflows with goroutine-driven concurrency, worker pools, context cancellation, and idempotent execution.

The product layer is built using TypeScript / Next.js, providing the interface where users monitor agents, review approvals, inspect audit trails, and manage workflows.

What You'll Own

• Design and own the agent loop (planning, tool dispatch, observations, retries, recovery)
• Build durable execution systems in Go with checkpointing and replayability
• Develop the TypeScript / Next.js governance and operations console
• Ensure idempotent and exactly-once execution semantics for tool actions
• Own context engineering, summarization, token budgeting, and structured outputs
• Build evaluation frameworks, regression suites, and trace-based testing systems
• Enhance governance capabilities including approvals, audit trails, validation, and security controls
• Review production code and mentor engineers across the team

You're Likely a Fit If You Have

• Built agentic AI systems from first principles, beyond framework orchestration
• Strong Go expertise for concurrent, stateful backend systems
• Strong TypeScript / Next.js experience
• 6+ years of experience building and operating production software
• Experience owning customer-facing multi-tenant SaaS platforms
• Strong distributed systems fundamentals (queues, retries, consistency, fault tolerance)
• Experience leading a module, platform area, or small engineering team

Nice to Have

• Experience building workflow or durable execution engines
• Direct integration with model-provider SDKs
• Agent security expertise (prompt injection defenses, tool sandboxing)
• Experience building AI evaluation, tracing, or guardrail tooling
• Deep AWS ownership (ECS, RDS, IAM, Terraform)
• 0→1 product or platform-building experience
• Strong observability practices (OpenTelemetry, tracing, cost attribution)

How We Work

Engineers who thrive here tend to:

• Prioritize correctness over cleverness
• Design for failure scenarios, not just happy paths
• Make assumptions explicit, especially when systems take real-world actions
• Use AI tools extensively while maintaining rigorous review standards
• Simplify systems without sacrificing reliability