About the Role
Join PwC's dynamic Agentic AI team as a Manager, where you will design, develop, and implement end-to-end agentic AI solutions. This role involves working with cutting-edge agentic frameworks, Retrieval-Augmented Generation (RAG), guardrail design, and full-stack implementations to drive operational efficiencies and innovation for our clients.
Responsibilities
- Gather business requirements and translate them into sophisticated agentic workflows.
- Develop and orchestrate AI agents using frameworks like LangChain, LangGraph, Google ADK, and AutoGen.
- Implement advanced RAG pipelines and memory-augmented workflows.
- Define structured data models using Pydantic.
- Implement robust guardrails using JSON Schema, regex, and LLM-as-judge techniques.
- Integrate APIs, enterprise applications, and MCP/A2A protocols.
- Build essential dashboards and microservices using FastAPI, Python, and ReactFlow.
- As a Manager, you will own the delivery roadmap, oversee multi-team orchestration, and engage with senior stakeholders.
Mandatory Skill Sets
- Strong proficiency in Python and FastAPI development.
- Experience with essential libraries such as Pandas and Pydantic.
- Hands-on experience with agent frameworks including LangChain, LangGraph, AutoGen, and Google ADK.
- Demonstrated expertise in RAG design and implementation.
- Proficiency in guardrail techniques such as JSON Schema, regex, and LLM-as-judge.
- Knowledge of CI/CD, containerization, and cloud-native deployment practices.
Preferred Skill Sets
- Knowledge of MCP server / A2A protocol.
- Experience with observability tooling like LangSmith and framework metrics.
- Full-stack development exposure, including React, ReactFlow, Next.js, and Node.js.
- Contributions to open-source projects are a plus.
Qualifications and Experience
- Experience: 12–15 years (Manager Level)
- Education: Bachelor's degree in Computer Science, Engineering, or AI/ML. Master's degree is preferred for senior levels.
- Required Qualifications: Bachelor of Technology, Master of Business Administration.