Sr. AI Engineer – Agent Orchestration

Averity; Pittsburgh (Allegheny County); On-site; USD 170,000 - 190,000; Full time

Benefits offered by this job

Unlimited PTO

401(k) with 4% vesting

Relocation assistance

Job summary

Averity is hiring a Senior AI Engineer focused on backend systems that orchestrate long-running AI agent workflows. You will design how an AI agent plans, sequences, and executes tasks, coordinating tools and data across sub-agents to automate work that once took hours.

The role emphasizes reliability, scalability, and human-in-the-loop safeguards, with a startup mindset and on-site work in Pittsburgh, PA. Prior experience in production-grade ML/LLM systems and security considerations is

Qualifications

• 5+ years of experience designing, building, and deploying production-grade ML/LLM-powered systems.
• Advanced Python skills are required.
• Hands-on AWS / cloud infrastructure experience.
• Familiarity with Git and modern development practices.
• Frontend framework knowledge (React, Next.js) is a plus.
• Active or obtainable US security clearance is a plus.

Responsibilities

• Architect and build the workflow engine that powers multi-step agent task execution with planning, sequencing, branching, and human handoff points.
• Design and implement multi-agent coordination patterns and tools orchestration across sub-agents and data sources.
• Build scalable tool-use infrastructure for dynamic tool discovery, selection, and chaining.
• Own reliability and observability for long-horizon tasks, including checkpointing and retry strategies.

Skills

LLM-powered systems design

Advanced Python

AWS / cloud infrastructure

Git

Frontend basics (React/Next.js)

Security clearance (US)

Job description

A high-growth technology company building AI-powered software for US government customers is hiring a Senior AI Engineer focused on agent workflow orchestration. The company's AI agent has seen rapid adoption, and the team is now pushing it to handle increasingly complex, multi-step tasks either fully autonomously or in a human-in-the-loop mode.

This role is squarely backend and systems-focused. You'll be designing how an AI agent plans, sequences, and executes long-running workflows — coordinating across multiple tools, data sources, and sub-agents to complete jobs that today require hours of manual work. Think of it as building the brain and nervous system of the agent, not the hands.

What You'll Do

• Architect and build the workflow engine that powers multi-step agent task execution — including planning, sequencing, branching, error recovery, and human handoff points
• Design and implement multi-agent coordination patterns: routing between specialized sub-agents, managing shared memory and context, and orchestrating parallel workstreams
• Build scalable tool-use infrastructure that lets agents dynamically discover, select, and chain tools based on task requirements
• Own reliability and observability for long-horizon agent tasks — designing checkpointing, retry strategies, and evaluation harnesses that catch failures before users do
• Define and enforce the technical architecture and engineering standards for agentic systems across the platform

What We're Looking For

• You have built an LLM reliably completing complex, multi-step tasks in production — not just single-turn Q&A, but real workflows with branching logic, tool calls, and failure handling
• Deep experience with workflow orchestration patterns (DAGs, state machines, event-driven architectures) applied to LLM-powered systems
• Hands-on experience with agent frameworks (Claude Agent SDK, OpenAI Agent SDK, LangGraph, or similar)
• Strong understanding of LLM reliability engineering: structured outputs, guardrails, token management, context window strategies, and evaluation pipelines
• Experience building systems that balance autonomy with human oversight (approval gates, escalation logic, confidence thresholds)
• Startup mindset — comfort with ambiguity, speed, and owning problems end-to-end

What Skills Do I Need?

• 5+ years of experience designing, building, and deploying production-grade ML or LLM-powered systems
• Advanced Python skills
• Hands-on AWS / cloud infrastructure experience
• Familiarity with Git and modern development practices
• Experience with modern frontend frameworks (React, Next.js) is a plus but not required
• Active or obtainable US security clearance is a plus

Compensation:

• Base Salary based upon experience but expect around $175,000
• Equity - between $125,000 - $175,000 with 4 year vest
• Unlimited PTO
• 401(k) with 4% immediate vesting

Please Note:

• You must be a US Citizen
• This is a Full-time, on-site (5 days) position in Pittsburgh, PA
• Role may include some travel - up to 10% - mainly between Pittsburgh and Washington DC area
• We offer relocation assistance if you are not currently local to the Pittsburgh area.