Agentic AI Architect

EXL • Jersey City, NJ

9h

About The Position

We are seeking an experienced Agentic AI Architect to design, develop, and deploy cutting-edge AI solutions. This role involves leveraging Generative AI and agentic frameworks to build intelligent systems that can reason, plan, and act. You will be responsible for integrating large language models (LLMs), implementing retrieval-augmented generation (RAG) techniques, ensuring AI safety and governance, and establishing robust evaluation and ModelOps practices. The ideal candidate will have a strong background in AI architecture, software development, and a deep understanding of the latest advancements in the AI field.

Requirements

• Expertise in GenAI & Agentic Frameworks (Semantic Kernel/LangGraph or similar orchestration frameworks).
• Proficiency in LLM integration (Azure OpenAI, OpenAI APIs, etc.).
• Strong skills in prompt engineering and prompt lifecycle design.
• Experience with Retrieval & RAG techniques (Azure AI Search, embedding pipelines, RAG design, grounding strategies, context management).
• Knowledge of Tool Access & Integration (MCP architecture, API design - FastAPI/REST/microservices, enterprise system integration).
• Familiarity with AI Safety & Governance tools (NVIDIA NeMo Guardrails, Microsoft Presidio).
• Experience with Evaluation & ModelOps (Azure AI Foundry, evaluation frameworks, prompt/version control).
• Proficiency in DevOps & Observability (CI/CD pipelines, logging, monitoring, performance tuning).
• Strong understanding of AI principles and architecture.

Nice To Haves

• Experience with specific enterprise systems and third-party APIs.
• Knowledge of PII detection/masking.
• Experience with prompt injection and hallucination control.
• Familiarity with App Insights.
• Experience with GitHub Actions.

Responsibilities

• Design and architect agentic AI systems using frameworks like Semantic Kernel or LangGraph.
• Integrate and manage LLMs through APIs such as Azure OpenAI and OpenAI.
• Develop and optimize retrieval-augmented generation (RAG) pipelines, including indexing, vector search, and embedding strategies.
• Implement AI safety and governance measures using tools like NVIDIA NeMo Guardrails and Microsoft Presidio.
• Establish and maintain ModelOps practices for evaluation, versioning, and monitoring of AI models.
• Design and implement tool access and integration mechanisms, including MCP architecture and API design.
• Ensure the scalability, performance, and observability of AI solutions through DevOps practices.
• Collaborate with cross-functional teams to define AI requirements and deliver impactful solutions.

What This Job Offers

• Job Type: Full-time
• Career Level: Senior
• Education Level: No Education Listed