Senior Software Engineer, AI Agent and Automation - USDS

Team Intro

We are the TikTok USDS Platform Responsibility - Automation Solutions team. Our mission is to build the foundational AI agent platform and intelligent automation workflows that safeguard the TikTok US ecosystem. Unlike traditional safety tools, we are moving towards an Agentic Future. We develop the "Agentic Service Bot" and "Multi-agent Orchestration" systems that empower Policy, SRE, and Compliance operations to handle complex risks (e.g. Election Safety, Minor Safety) at scale. We bridge the gap between advanced ML models and production-grade operational excellence in the secure TTP environment.

Responsibilities

• Own the end-to-end technical roadmap for AI agent workflows or platforms within the USDS safety domain.
• Design and build core platform components including agent runtime, state management, and tool-calling frameworks to enable multi-agent coordination.
• Build modern UIs and workflow editors that allow operations teams to monitor, debug, and configure complex AI workflows.
• Partner with Data Science teams to integrate LLMs and detection models into reusable workflow nodes with high reliability and low latency.
• Drive the migration of critical data pipelines to USDS and ensure 100% recovery for safety-critical services via robust system design.
• Collaborate closely with product team to identify high-impact automation scenarios and translate ambiguous safety needs into scalable engineering solutions.

Qualifications

Minimum Qualifications:

• Bachelor's or Master's degree in Computer Science, or a related technical field.
• 5+ years of professional software engineering experience building large-scale production systems.
• Strong backend experience in Go (preferred), Java, or Python, with a deep understanding of distributed systems, concurrency, and microservices architecture.
• Hands-on experience with modern web technologies (e.g. TypeScript, React), with the ability to build complex, interactive internal tools.
• Practical experience integrating LLM/ML capabilities into products (e.g. prompt engineering, handling model latency, evaluation flows).
• Prior experience in a 0→1 or early-stage builder role, with a track record of shipping ambiguous projects.

Preferred Qualifications:

• Experience building workflow engines (e.g. Temporal, Airflow) or node-based editors (e.g. React Flow).
• Familiarity with applied LLM patterns like RAG, Multi-step Agents, and Tool-calling.
• Prior experience in Trust & Safety, Privacy, or Regulatory engineering is a plus.