Roku
Software Engineer, Embedded Agentic AI
About this role
Roku seeks a hands-on Agentic AI Engineer to design, build, and deploy intelligent agents and copilots for Roku TV embedded systems. You'll own the full lifecycle from prototyping through production, architecting durable AI systems that operate reliably in resource-constrained environments while integrating with internal tools and services.
What you'll do
- Architect and deploy AI agents and copilots for Roku TV, integrating with internal systems and services
- Own end-to-end agentic systems including model selection, prompt design, RAG pipelines, and conversational interfaces
- Design single-agent and multi-agent orchestration patterns with handoffs, delegation, and cooperative execution
- Implement tool-calling and MCP-style integrations enabling agents to safely interact with surrounding systems
- Build evaluation, observability, and monitoring for conversation quality, task success, latency, and cost metrics
- Create safeguards for production readiness including testing pipelines, controlled rollouts, and drift detection
What they're looking for
- Python and C/C++ or systems programming languages
- LLM-based systems and prompt engineering
- RAG pipeline design and implementation
- Agent orchestration frameworks and multi-agent patterns
- Backend development and production systems design
- Observability and monitoring implementation
- Model evaluation and quality assessment
- Software engineering fundamentals and testing practices
Benefits
- Health insurance
- Equity awards
- Life insurance and disability benefits
- Parental leave
- Wellness benefits
- Paid time off
Opens the official application on the employer’s site. No login required.
Roku
Roku builds a TV streaming platform enhanced by AI and machine learning to deliver high-quality audio, video, and user experiences to millions of global viewers. The company is hiring Software Engineers and Machine Learning Engineers to develop embedded systems, intelligent AI agents, and engaging platform features across its streaming media and UI teams.
- Website
- roku.com
Likely interview questions
- Describe your experience designing and deploying LLM-based systems—what challenges did you encounter with prompt engineering or retrieval quality in production?
- How have you approached building RAG pipelines? What techniques did you use to keep them aligned with evolving data sources?