Agentic AI Engineer

Agentic AI System Design Design and implement multi‑agent architectures including: Orchestrator / supervisor agents Task‑specialized agents (research, extraction, validation, decisioning) Reflection, critique, and self‑correction loops

Build goal‑driven agent workflows with constrained autonomy and human‑in‑the‑loop patterns. Define agent boundaries, decision policies, and escalation logic. LLM, RAG & Tooling Build enterprise‑grade RAG systems : Ingestion, chunking, metadata enrichment Vector indexing and retrieval strategies Grounded generation and citation control

Implement tool‑calling agents that interact with: APIs, databases, search systems Internal platforms and workflows

Optimize latency, cost, and accuracy across LLM interactions. Production Engineering & AgentOps Build agentic systems as scalable backend services (FastAPI / REST / async services). Apply strong software engineering discipline : Modular code, clean abstractions, testability CI/CD, versioning of prompts, tools, and agents

Implement AgentOps / LLMOps , including: Evaluation harnesses (prompt, retrieval, agent behavior) Observability (traces, metrics, decision paths) Rollback and controlled rollout strategies

Ensure robustness against hallucinations, loops, tool failures, and unsafe actions. Governance, Safety & Reliability Implement guardrails, policies, and monitoring for agent behavior. Design systems with traceability, auditability, and explainability . Ensure secure handling of sensitive data (PII/PHI where applicable). Enforce responsible‑AI principles in autonomous systems. Technical Leadership Lead design reviews and mentor junior engineers. Influence platform standards for agentic AI across teams. Partner with product managers, architects, and domain SMEs to translate workflows into agent behavior.