Applied AI Engineer, Sarvam Agents

About Sarvam Sarvam is building the bedrock of Sovereign AI for India. The company is developing India’s full-stack sovereign AI platform, building across research, models, infrastructure and applications with a singular focus on making AI genuinely work for India. Sarvam works with leading enterprises and public institutions and is backed by Lightspeed, Peak XV, and Khosla Ventures. Sarvam partners with India’s leading brands, including Tata Capital, SBI Life, CRED, IDFC, and LIC. About the Role Sarvam Agents bring AI into the way teams actually work. They plug into the tools and workflows of an organization (email, docs, calendars, internal systems, the long tail of SaaS), take multi-step action on a user's behalf, run on a schedule when no one is watching, and get better as the team uses them. Shared across the org, scoped by the right permissions, observable by admins. This is what agents look like when they have to actually work. We're hiring an Applied AI Engineer to build these agents and own them in production. You'll design the flows, wire up the tools, ship the harness, and be on the hook when something breaks. What You’ll Do - Build agents end-to-end: flow design, prompting, tool definitions, memory, evaluation, deployment - Own agent reliability in production: debugging real failures, designing guardrails, running incident response when things break - The agent runtime: state management, retries, scheduled triggers, long-running execution that keeps moving when no one is watching - Connector and integration framework: OAuth, third-party SaaS APIs, the plumbing that lets agents reach into real workplaces - MCP server infrastructure at scale: the tool layer agents reach for - Memory and context engineering: knowing when to use working memory vs. long-term memory vs. retrieval, and building the systems that make it work (Honcho and friends) - Multi-tenancy and sharing: agents that live inside organizations, with the right permissions, audit trails, and admin surfaces - Evals: the test infrastructure that tells us whether a change made the agent better or worse, before it touches a user - Tool design discipline: schemas, error messages, idempotency, retry semantics. Tools your agents actually use well. - Streaming, cost, and latency engineering for live agent traffic - Observability and integration testing for agent runs - Prompt iteration workflows: versioning, A/B testing, safe rollback What We’re Looking For - 3 to 5 years of total experience, with 2+ years of strong backend development experience in Python - LLM API fluency: prompt design, structured outputs, function calling, working with major model APIs at production quality - Agentic framework experience: LangGraph, ADK, or equivalent - Production experience running MCP servers at scale - Shipped AI features to real users in production, not just notebooks or prototypes - RAG patterns: chunking, embeddings, hybrid search, reranking, vector databases - Building and running evals for AI systems - Cost and latency optimization for LLM workloads - Observability fluency: Signoz, Datadog, or any equivalent stack - Experience integrating with third-party SaaS APIs and OAuth flows - PostgreSQL or MySQl, Redis Bonus Points - Open-source contributions in the agent / LLM / harness ecosystem - Familiarity with eval frameworks (LangSmith, Braintrust, and similar) - Experience with human-in-the-loop / approval flow design for AI systems - Time at an early- or growth-stage startup: comfort with ambiguity, speed, and ownership Why Sarvam? Sarvam is a fast-moving, high talent-density team building full-stack AI for India, working on problems that push the frontiers of AI with real population-scale impact. - Work alongside researchers, engineers, builders, and business leaders who move fast and hold each other to a very high bar - High ownership and high impact, from day one - Everything we do is AI-first, from the way we build and ship to the way we think about problems - You ca