Lexsi Labs (AryaXAI), a frontier AI lab building aligned, interpretable and safe AI systems, is hiring remote AI Agent Engineer Interns for its Applied AI team. You will work on real agentic systems in one or more tracks. In applied agentic AI, you will build enterprise agents that reason over workflows, use tools and data, and run multi-step processes reliably. In agent R&D and evaluation, you will work on harnesses, benchmarks, long-horizon reasoning, failure analysis, alignment and interpretability, which may lead to papers. In product and tooling, you will build agents for alignment workflows, AI-scientist-style experimentation and debugging. You will turn open-ended ideas into prototypes, experiments and research artifacts, document your findings, and work with research, engineering and product teams. Candidates need strong programming skills, ideally in Python, hands-on experience with LLMs, and a real interest in agents, evaluation, alignment or enterprise AI. Bonus: agent frameworks, evaluation pipelines, open-source work, backend systems or AI safety. The listing marks this role as part-time.