Agent Engineer

About this job

As an Agent Engineer at Sarvam AI in Bengaluru, you will build and deploy production-grade AI agents for India's largest banks, insurers, NBFCs, government departments, and fast-growing companies. Sarvam is a leading full-stack AI company building sovereign AI infrastructure, foundational models, and voice-first AI agents for millions of people. A normal day in this full-time role involves designing agent workflows, integrating AI systems with complex enterprise software like core banking and CRMs, building robust evaluation pipelines, and engineering agents like versioned software rather than simple console prompts. You will work alongside leading AI researchers and engineers to ensure systems handle high-volume real interactions with reliability, accuracy, and cost-efficiency at scale.

Responsibilities

• Build production agents on the Sarvam stack including prompts, system design, workflows, tool integrations, and evals.
• Map agent scenarios and edges to ensure comprehensive design coverage rather than discovering issues in production.
• Build fast proofs of concept for customer validation before scaling.
• Integrate AI agents into complex legacy systems such as core banking, CRM, ticketing, and telephony.
• Own end-to-end quality by defining metrics, building evaluation pipelines, and tuning for accuracy, cost, and reliability.
• Treat agents as engineered software by implementing versioning, code reviews, and regression suites.
• Scale patterns efficiently across new customer environments while maintaining high performance under real load.

Requirements

• 3+ years of experience with a proven track record of shipping systems to production and maintaining them.
• Serious experience working with LLMs and agents, including context engineering, tool use, structured outputs, memory, and failure mode handling.
• Experience using agentic frameworks alongside both open and closed source models.
• Strong engineering discipline focused on versions, tests, regressions, reproducibility, and robust evaluation metrics.
• Strong Python skills with broad technical range across backend services, data pipelines, APIs, and cloud infrastructure.
• Ability to design scalable systems that maintain consistent behavior over months of real load and usage.
• Strong communication skills to collaborate directly with customer engineering teams and understand technical constraints.

Who should apply

This role suits experienced software and AI engineers with 3+ years of professional experience who have moved past simple prototypes and successfully shipped real-world LLM or agent systems that users depend on. Candidates should have a strong background in Python, rigorous engineering discipline, and experience handling failure modes at scale. If you enjoy solving complex deployment challenges for major enterprises and government bodies while working directly on a full-stack AI platform, this position is ideal for you. Freshers cannot apply as the minimum requirement is 3+ years of experience.

Frequently asked questions