Agentic AI Orchestrator

Job Title: Agentic AI Orchestrator

Location: Remote | Salary: Full-Time | $140K–$170K Base + Performance Bonus (Total Comp Potential ~$220K) | Reports to: CEO | Job #: 7372

About the role
This is one of the newest and fastest-growing roles in AI. As an Agentic AI Orchestrator, you design and operate systems where multiple specialized AI agents collaborate to complete complex, multi-step workflows — autonomously and reliably in production.

You're not building chatbots. You're architecting coordination layers: defining how agents communicate, delegate, fail gracefully, and recover — at scale. This role sits at the intersection of systems engineering, AI architecture, and product thinking.

What you'll architect

Agent coordination: Define communication protocols between specialized agents

Task routing: Distribute subtasks to the right agent at the right time

State & memory: Manage shared context, working memory, and checkpointing

Failure handling: Design fallback logic, retries, and graceful degradation

What you'll own

Design end-to-end multi-agent architectures — from orchestration topology to inter-agent communication schemas

Build and maintain the orchestration layer using frameworks such as LangGraph, AutoGen, CrewAI, or custom-built systems

Define agent roles, capabilities, and boundaries — research agents, coding agents, data agents, tool-use agents, and critic/evaluator agents

Implement robust task decomposition: break complex goals into agent-executable subtasks with clear success criteria

Manage shared memory, vector stores, and context windows across agent interactions to prevent hallucination and state drift

Design tool-use layers: connect agents to APIs, databases, code interpreters, web search, and internal systems

Build observability into multi-agent pipelines — trace agent calls, log decisions, monitor token costs, and detect failure patterns

Establish evaluation frameworks for agentic workflows: correctness, latency, cost per task, and task completion rates

Collaborate with ML Engineers and product teams to translate business workflows into agent-executable pipelines

Maintain security and permission boundaries — agents must not exceed their defined scope or access unauthorized systems

Required

3+ years software or AI engineering, systems in production

Python (expert) — async patterns, concurrency

LLM APIs: OpenAI, Anthropic, or Gemini (tool use, function calling)

Agent frameworks: LangGraph, AutoGen, or CrewAI

RAG pipelines and vector database integration

REST / GraphQL API design and consumption

Prompt engineering for agent personas, instructions, and constraints

Strong debugging instinct — agents fail in non-obvious ways

Strong plus

Custom orchestration without a framework (raw LLM + state machine)

Multi-modal agents: vision, code execution, document processing

LangSmith, Arize, or custom eval tooling for agents

Experience with OpenAI Assistants API or Anthropic tool use

Fine-tuned models used as specialized sub-agents

Workflow automation (n8n, Zapier, or custom task queues)

Security design for agentic systems (scope, sandboxing, guardrails)

Experience in B2B SaaS, recruiting tech, or operations automation

Tech stack
Python, LangGraph, AutoGen, CrewAI, OpenAI / Anthropic APIs, Pinecone / pgvector, Redis, FastAPI

#J-18808-Ljbffr