About the role

Freelance Software Engineer (Remote)

Fully remote (United States) — full-time freelance.

Role Overview

Help advance autonomous, agent-based AI systems by training and evaluating LLMs so they can act as proactive, multi-step agents. You’ll also build production-grade software components and provide expert feedback to improve AI performance, reliability, and safety.

Responsibilities

Train and evaluate AI agents to interpret, reason, and execute complex multi-stage workflows.
Build and assess production-grade software with modular architectures (e.g., data parsing, logic processing, orchestration, reporting).
Review and analyze multi-turn system interactions to identify strengths, gaps, and failure modes.
Provide clear, structured, high-quality technical feedback to improve AI outcomes (performance, reliability, safety).
Support integrations with live tools, databases, and external APIs to enable real-world problem solving.

Requirements

2+ years of experience in backend engineering, AI automation, software development, or complex systems integration.
Strong proficiency in at least two programming languages (e.g., Python, JavaScript, Go, Java).
Experience with SQL databases and live production (non-mocked) environments.
Hands-on experience managing stateful, multi-turn system interactions.
Strong attention to detail and ability to deliver concise, high-density technical feedback.

Nice to Have

Experience building multi-stage coordination/orchestration workflows using data-informed reasoning.
Experience integrating systems with live tools and APIs (databases, email platforms, third-party services).
Comfort implementing persistent state and session tracking for long-running tasks.
Experience identifying subtle system risks (privacy leaks, authority escalation, indirect prompt injection).

About SoTalent

SoTalent is a staffing/recruiting firm that matches talent with clients building advanced technology products. In this role, the client is focused on developing next-generation autonomous, agent-based AI systems, including training and evaluation of large language models (LLMs) for real-world workflows.

Tags