Full-Stack Software Engineer

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

Job summary

Syntic AI, Inc. is seeking an experienced backend-focused engineer to build the runtime that enables AI agents to plan tasks, use tools, retain relevant memory, and recover from failures.

The role focuses on agent orchestration and the backend systems that make agent behavior effective, observable, interruptible, and safe. You will design and implement systems coordinating agent workflows, ensure robust task execution, and improve safety and observability across distributed components.

Qualifications

• Experience building LLM-powered applications or agents for production use.
• Strong backend engineering skills in TypeScript or Python.
• Understanding of failure modes in models that use tools.

Responsibilities

• Design and build systems for coordinating agent workflows and tool calls.
• Improve task completion through measurable evaluations.
• Make agent behavior observable so its execution can be understood.
• Support interruption and recovery in agent workflows.
• Address safety concerns in tool-using agent behavior.

Skills

LLM-powered apps

Backend engineering

TypeScript

Python

Observability

Tool-failure safety

Job description

Role overview

Build the runtime that enables AI agents to plan tasks, use tools, retain relevant memory, and recover from failures. The work centers on agent orchestration and the backend systems that make agent behavior effective, observable, interruptible, and safe.

Responsibilities

• Design and build systems for coordinating agent workflows and tool calls.
• Improve task completion through measurable evaluations.
• Make agent behavior observable so its execution can be understood.
• Support interruption and recovery in agent workflows.
• Address safety concerns in tool-using agent behavior.

Requirements

• Experience building LLM-powered applications or agents for production use.
• Strong backend engineering skills in TypeScript or Python.
• Understanding of failure modes in models that use tools.