AI Orchestration & Retrieval Architect

Vertex, Inc.; Pennsylvania; Hybrid; USD 159,600 - 207,500; Full time

Job summary

Vertex, Inc. seeks a Principal Engineer to design the central AI system, connecting LLMs to tools, data, and sub-agents. You will shape retrieval, chunking, and how tools are surfaced to models, including MCP servers, with a hands-on approach.

You will mentor engineers, define multi-agent patterns, and balance cost, latency, and safety while collaborating with product teams to onboard capabilities as tools and agents.

Qualifications

• Bachelor’s degree in Computer Science, Engineering, or related discipline; 12+ years of software/AI engineering experience.
• Hands-on experience building LLM orchestration, agents, and retrieval systems.
• Ability to set strategy and standards while remaining hands-on in code.

Responsibilities

• Design orchestration and abstraction layers connecting LLMs to tools, data, and sub-agents.
• Build and operate MCP servers; define tool exposure and versioning standards.
• Define tool-surface strategy and multi-agent patterns; determine when to use sub-agents vs direct exposure.
• Design retrieval systems (RAG): chunking, embeddings, vector stores, and re-ranking.
• Establish observability and tracing across multi-step agent/tool calls; manage safety and access control.
• Partner with product teams to onboard capabilities as tools and agents; mentor engineers.

Skills

LLM orchestration

Agentic patterns

Observability

Stakeholder collaboration

Cost/latency optimization

Context-window management

Retrieval/RAG

System design

Education

Bachelor's degree in Computer Science or related field

Advanced degree preferred

Tools

LangGraph

LlamaIndex

Semantic Kernel

MCP servers

Job description

Vertex, Inc. seeks a Principal Engineer to design the central AI system, connecting LLMs to tools, data, and sub-agents. You will shape retrieval, chunking, and how tools are surfaced to models, including MCP servers, with a hands-on approach.

You will mentor engineers, define multi-agent patterns, and balance cost, latency, and safety while collaborating with product teams to onboard capabilities as tools and agents.