Principal AI Engineer

About the role

Looking for a Staff/Principal AI Engineer to lead enterprise‑scale Agentic AI implementations for Fortune 500 clients. This is a highly hands‑on engineering role focused on designing and deploying autonomous, tool‑calling AI systems that operate reliably at enterprise scale.

Responsibilities

• Lead the architecture, design, and production deployment of enterprise‑scale Agentic AI systems.
• Develop autonomous, tool‑calling AI agents that integrate with APIs, microservices, and distributed systems.
• Implement production‑grade data pipelines, RAG pipelines, vector retrieval, and hybrid retrieval solutions.
• Design and enforce guardrails, validators, policy checks, and self‑correction mechanisms.
• Optimize token usage, context windows, latency, and cost through caching, model routing, batching, and streaming.
• Establish observability, traceability, and regression detection using LLM evaluation frameworks.
• Own technical roadmaps and communicate architecture decisions to client stakeholders.
• Mentor and provide technical leadership to engineering teams.
• Navigate ambiguous environments and drive solutions from concept to production.

Requirements

• 15+ years of professional software engineering experience.
• 2+ years of hands‑on LLM engineering experience, with proven production deployments of Agentic AI systems.
• Active PMP certification.
• Strong ability to work effectively in ambiguous, fast‑moving environments.

Skills

• Python (strong hands‑on development).
• Backend/system languages: Go, Rust, Java, C/C++.
• Data structures, algorithms, and system design.
• APIs, microservices, and distributed systems.
• Production‑grade data pipelines.
• Agentic AI / LLM frameworks: LangGraph, Google ADK, CrewAI, Claude Agent SDK or equivalents.
• MCP servers & tool calling.
• RAG: chunking, embeddings, vector/hybrid retrieval, reranking.
• Vector databases: Pinecone, Milvus, Weaviate, FAISS.
• Guardrails, validators, policy checks, self‑correction, deterministic fallbacks, circuit breakers, rollback strategies.
• Optimization & evaluation: token/context‑window optimization, latency & cost optimization, caching, model routing, batching, streaming.
• LLM evaluation frameworks & regression detection: LangSmith, Langfuse.
• Production observability & traceability.
• Cloud & DevOps: AWS (ECS/EKS, Lambda, S3, DynamoDB, Redshift, Step Functions); Azure/GCP experience a plus.
• CI/CD & infrastructure as code: Terraform, CloudFormation.
• Containerization and deployment: Docker.

Schedule

EST overlap required.

Location

New York City, NY (on‑site 3 days per week).

Employment type

Full‑time, on‑site (3 days WFO).

Contact

Submit your resume to rachael@sidramtech.com.