AI Agent – Multi‑Agent System Engineer

About the role

We are looking for an experienced AI Agent specialist to design, build, and operate sophisticated multi‑agent systems. The role is based in Pune and requires hands‑on delivery of agentic workflows for real‑world products.

Key responsibilities

• Design and implement multi‑agent architectures with planners, workers, and specialized sub‑agents, including hand‑off protocols and failure handling.
• Develop and maintain MCP servers that expose internal APIs, databases, and workflows securely and observably.
• Integrate MCP clients using transports such as STDIO/HTTP, handling authentication, authorization, and multi‑server connections.
• Engineer context management pipelines (chunking, summarization, windowing) and choose between MCP tools, private stores, or RAG/vector search based on latency, security, and governance.
• Create robust prompts and safety guardrails to mitigate injection, stuffing, and data‑exfiltration risks.
• Set up autonomous coding loops (Ralph‑style) with test, linter, and stop‑condition gating, and monitor logs, metrics, and traces.
• Engage with customers to gather requirements, present proposals, and maintain regular communication.
• Leverage cloud platforms, especially AWS, to deploy and scale solutions.

Required profile

• 10+ years of overall experience with at least 2 years focused on agentic AI systems.
• Proven ability to translate complex technical concepts into clear explanations for engineers and stakeholders.
• Strong customer‑facing communication skills and a pragmatic approach to balancing quality, safety, and delivery speed.

Required skills

• Multi‑agent system design
• MCP server development
• MCP client integration (STDIO/HTTP)
• Authentication & authorization mechanisms
• Context engineering (chunking, summarization, windowing)
• RAG / vector search
• Prompt engineering and safety guardrails
• Ralph‑style autonomous workflow loops
• Observability, logging, and debugging of agent runs
• AWS cloud services