AI Systems Architect

About the Role

We’re building systems that don’t just use AI models — they learn, adapt, and optimize themselves in production . We’re looking for an AI Systems Architect to design and evolve the technical backbone behind large-scale, feedback-driven AI systems.

This role sits at the intersection of LLM infrastructure, evaluation systems, optimization workflows, and production reliability . You’ll help define how AI systems observe themselves, learn from real-world signals, and improve over time — with engineers staying firmly in control.

This is not a research-only role and not a UI-heavy role. It’s about system design, feedback loops, and production-grade intelligence .

What You’ll Do

• Architect end-to-end AI systems that operate in production (LLMs, agents, evaluators, optimizers).
• Design feedback loops that turn metrics, logs, and human feedback into measurable system improvements.
• Define evaluation strategies for model quality, cost, latency, and regression prevention.
• Lead optimization workflows such as prompt evolution, fine-tuning, model routing, distillation, or reinforcement learning.
• Build or guide infrastructure for observability, experimentation (A/B testing), and automated rollouts.
• Partner with product and engineering teams to translate real-world problems into scalable AI architectures.
• Make principled tradeoffs between performance, reliability, cost, and speed.

Requirements

• Strong systems thinker with experience designing complex, production AI or ML systems .
• Hands-on experience with LLMs, agents, or ML pipelines beyond simple API usage.
• Deep understanding of evaluation, experimentation, and optimization in real environments.
• Comfort working across backend systems, data pipelines, and model interfaces.
• Ability to reason clearly about failure modes, edge cases, and system incentives.
• Strong communication skills — you can explain why a system works, not just how .

Nice to Have (Not Required)

• Experience with reinforcement learning, model fine-tuning, or distillation.
• Prior work on open-source infrastructure or widely used internal platforms.
• Background in distributed systems, compilers, or performance-critical software.
• Experience designing AI systems used by external customers at scale.