AI Backend Engineer

As an AI Backend Engineer at Zig.ai in Bangalore, you will be instrumental in owning and developing core backend and infrastructure systems that power our advanced orchestration layer. This role involves designing and implementing stateless backend services that operate over durable workflow states, with a strong emphasis on explicit workflow and execution state models, including statuses, transitions, and events. You will build event-driven orchestration systems that incorporate robust retry mechanisms, idempotency, and recovery protocols. The systems you develop will power AI-driven execution, going beyond mere recommendations.

We are deeply committed to correctness, observability, and enabling fast iteration. You will be responsible for designing essential primitives, defining state models, and rapidly deploying to production. Close collaboration with customers is a key aspect of this role, allowing you to understand real-world workflows, identify edge cases, and diagnose failure modes. You will work daily with our founders, frontend engineers, product team, and directly with customers to ensure our solutions are effective and reliable in production environments.

Key responsibilities include designing and building stateless backend services using Python and FastAPI, defining durable workflow state models based on customer workflows, and integrating AI agents into reliable execution pipelines. You will also build event-driven asynchronous systems such as queues, schedulers, and background jobs, own prompt engineering, and manage production LLM integrations. Designing clean APIs for consumption by frontend systems and agents is also a critical part of the role.

You will engage directly with customers to understand workflow failures, debug production issues, and rapidly iterate based on feedback, ensuring our systems are observable, debuggable, and resilient. You will operate backend and AI infrastructure end-to-end.

Ideal candidates will possess strong backend experience (5+ years preferred) with deep proficiency in Python and backend systems. Experience modeling durable, explicit state for asynchronous workflows is essential, as is practical experience with LLMs and prompt engineering in production. You should be comfortable engaging directly with customers and reasoning about complex real-world workflows. The ability to translate ambiguous problems into clean system designs is highly valued. A startup mindset, characterized by fast iteration, comfort with ambiguity, and a high degree of ownership, is crucial. Desirable, but not required, skills include experience with workflow engines (Temporal, Prefect, n8n, LangGraph, Agno), event sourcing, open-source LLMs, and B2B SaaS or sales-tech experience.

This role requires hands-on experience with Docker and containerized services, Kubernetes (EKS/ECS or equivalent), CI/CD pipelines, AWS and/or GCP cloud platforms, and implementing comprehensive logging, metrics, tracing, and alerting solutions.