AI Engineer at Ruya AI

on-site · full-time · Visa sponsorship

Apply for this role at Ruya AI

Responsibilities:
- Design and implement agent orchestration, subagent delegation, tool-calling, and structured output pipelines for multi-modal intelligence agents.
- Build asynchronous services using Python/TypeScript with FastAPI, Pydantic, and asyncio to operate reliably at scale.
- Develop LLM pipelines: prompt/context management, memory and token strategies, and safe tool integrations.
- Implement safeguards against prompt injection, data leakage, and unsafe tool behavior.
- Create behavioral evaluations, regression datasets, and tracing for model calls and tool execution (Langfuse, Sentry, OpenTelemetry).
- Diagnose failures across models, tools, data sources, and distributed services; collaborate with AI and engineering teams to iterate.

Requirements:
- Strong production experience in Python and asynchronous programming; experience with FastAPI/Pydantic preferred.
- Practical experience building LLM applications that use tools and multi-step reasoning.
- Familiarity with structured output validation, non-deterministic system testing, and empirical evaluation of AI behavior.
- Understanding of prompt injection risks, sensitive-data exposure, and mitigation strategies.
- Ability to trace, measure, and debug model and orchestration failures across distributed systems.
- Preferred: experience with MCP, Langfuse, LiteLLM/Anthropic-compatible APIs, agent memory/graph systems, geospatial intelligence, or model fine-tuning/RAG techniques.