AI Inference & Optimization Engineer

Syntic AI, Inc.; United States; Remote; USD 120,000 - 180,000; Full time

Job summary

Syntic AI, Inc. in the United States is seeking a backend engineer to develop the runtime that supports AI agents as they plan work, call tools, use memory, and recover from errors.

You will build systems for agent orchestration and tool calling, and improve inference outcomes through measurable task evaluations. Focus is on making agent behavior observable and safe, supporting planning, memory, and recovery in the runtime.

Qualifications

• Experience building production LLM applications or agents.
• Strong backend engineering skills in TypeScript or Python.
• Understanding of failure modes in tool-using models.

Responsibilities

• Build systems for agent orchestration and tool calling.
• Support agent planning, tool use, memory, and recovery in the runtime.
• Use measurable evaluations to identify and improve task success rates.
• Improve visibility into agent behavior and support safe interruption.
• Account for failure modes in models that use tools.

Skills

Backend engineering

TypeScript

Python

LLM applications

Agent orchestration

Job description

Role overview

Develop the backend runtime that supports AI agents as they plan work, call tools, use memory, and recover from errors. The work focuses on agent orchestration and tool use, with an emphasis on improving inference outcomes through measurable task evaluations and making behavior observable and safe.

Responsibilities

• Build systems for agent orchestration and tool calling.
• Support agent planning, tool use, memory, and recovery in the runtime.
• Use measurable evaluations to identify and improve task success rates.
• Improve visibility into agent behavior and support safe interruption.
• Account for failure modes in models that use tools.

Requirements

• Experience building production LLM applications or agents.
• Strong backend engineering skills in TypeScript or Python.
• Understanding of failure modes in tool-using models.