Artificial Intelligence Engineer (AWS Bedrock)

Job Description

Role : AI Engineer

Location - Pune
Experience - 8 - 10 years

Key Responsibilities

• Build the agent's grounded RAG pipeline on Bedrock Knowledge Bases, including prompt/instruction engineering and response grounding with citations
• Implement Bedrock Guardrails enforcement for read-only scope restriction and PII redaction, and integrate the agent with its chat channel
• Curate and help build the internal library of reusable MCP servers and connectors (Library & MCP Servers) and the Loom pattern/prompt library packaged for the IDE
• Design and build the Context Serving API's agent-facing integration and configure agent-specific memory strategies on top of platform Memory
• Build all five R3.0 quick-follow agents (Invoice & AP, Learning & Development, Contract Lifecycle, Store Task Orchestration, Indirect Spend) by reusing the proven orchestration pattern, wiring each into Context Serving, its Domain Graph and the Retail Ontology rather than building bespoke integrations
• Run each new agent through the shared Evaluation framework instead of building a bespoke test harness per agent

Required Qualifications

• 7-10 years in AI, with hands-on experience building LLM-based or agentic applications in production, preferably on AWS Bedrock and AgentCore Platform
• Practical experience with retrieval-augmented generation (RAG): chunking, embeddings, grounding and citation strategies.
• Experience with prompt/instruction engineering and iterating on agent behavior against evaluation criteria.
• Experience integrating agents with external systems via typed/schema-validated tool calls or function calling (MCP experience a strong plus).
• Comfortable working across multiple agent codebases with shared conventions this role is expected to reuse platform patterns rather than rebuild them per agent.

Preferred Qualifications

• Experience with AWS Bedrock Agents, Guardrails and Knowledge Bases specifically.
• Familiarity with building or maintaining internal developer tooling (snippet libraries, MCP servers, IDE extensions)
• Exposure to HR, ITSM, finance/AP or procurement domain workflows.

Email your updated profiles on [email removed]