Role: AI Engineer – Agentic AI / AWS Bedrock

Experience: 7+ Years

Location: Chennai, Bangalore, Pune

Work Model: Hybrid – 3 Days WFO

Payroll: Direct Payroll – Haparz

Budget: Up to 30 LPA

Role Overview

We are looking for an experienced AI Engineer who can understand a manual business process, convert it into an AI-agent workflow, build and deploy the solution on AWS, evaluate its performance, and continuously improve it in a production setting.

This role requires strong hands-on engineering experience across Agentic AI, AWS Bedrock AgentCore, Step Functions, Claude, MCP, Python, RAG, observability and evaluation. The candidate should be comfortable owning solutions end-to-end and working directly with enterprise clients and non-technical stakeholders.

What You’ll Do

- Analyze manual business processes and identify opportunities to automate them using AI agents.
- Design agent harnesses covering control loops, tools, skills, context, state management and human-in-the-loop checkpoints.
- Build deterministic and LLM-driven workflows with clear policy gates, exception handling, timeout controls and loop guards.
- Develop AWS Step Functions workflows using Amazon States Language (ASL) for branching, retries, parallel execution and human approval flows.
- Integrate Step Functions with Amazon Bedrock AgentCore Runtime, Lambda and other AWS services.
- Build, deploy, operate and scale agent workloads using Amazon Bedrock AgentCore.
- Implement AgentCore Gateway and expose enterprise tools and services securely to agents.
- Build reusable agent skills and integrate them into production workflows.
- Develop and consume MCP servers to provide agents with controlled access to enterprise data and tools.




- Implement multi-agent workflows involving task delegation, coordination and result aggregation.
- Build RAG pipelines using appropriate chunking, embeddings, retrieval and vector database strategies.
- Develop production-grade Python services using Python 3.10+, async/await and type hints.
- Build backend services using FastAPI or equivalent frameworks and expose secure REST APIs.
- Implement authentication and authorization using JWT/OAuth2.
- Work with SQL, NoSQL and vector databases including DynamoDB, pgvector, Pinecone or OpenSearch.
- Build document-processing workflows for PDFs, Excel files and unstructured business data.
- Implement AgentCore Observability with tracing, metrics, logging and step-level audit trails.
- Monitor agent performance using tools such as CloudWatch and Datadog.
- Design and execute agent evaluations using AgentCore Evaluations and golden datasets.
- Establish regression testing to measure agent quality before and after production releases.
- Configure Amazon Bedrock Guardrails for prompt-injection defence, PII protection, hallucination mitigation and safety controls.
- Build agent-facing interfaces using AG-UI, including streaming responses, tool-call visibility, shared state and human approval interactions.
- Use Claude Code and/or Cowork to design, develop and scale AI-powered production workflows.




- Work directly with clients to understand requirements, present technical solutions and translate business processes into AI-powered solutions.
- Own solutions from discovery and architecture through development, deployment, evaluation and continuous improvement.

What We’re Looking For
- Strong software engineering background with substantial hands-on AI engineering delivered to production.
- Proven experience building and deploying LLM-agent or Agentic AI solutions in production.
- Strong understanding of agent architecture, including harness design, tool use, context, state, control loops and human-in-the-loop mechanisms.
- Hands-on experience with AWS-based AI workloads, preferably Amazon Bedrock AgentCore and AWS Step Functions.
- Strong Python development experience, including asynchronous programming and production API development.
- Experience with Claude, Claude Code and/or Cowork in enterprise AI workflows.
- Hands-on knowledge of MCP, tool integration, skills and multi-agent architectures.
- Experience building and evaluating RAG solutions.
- Understanding of agent observability, evaluation, guardrails and production monitoring.
- Experience with FastAPI, REST APIs, SQL and NoSQL databases.
- Experience with vector databases such as pgvector, Pinecone or OpenSearch.
- Ability to convert ambiguous business requirements into practical AI solutions.
- Strong ownership mindset with the ability to work independently with minimal guidance.
- Excellent communication skills and experience working directly with enterprise customers and non-technical stakeholders.
- Consulting or client-facing AI engineering experience will be an advantage.

Work Location: In person

📌 Hiring AI Engineer – Agentic AI / AWS Bedrock (Chennai)
🏢 Haparz
📍 Chennai