Staff Software Engineer, Agentic, Optimization, Recursive Self-Improvement

Google's software engineers develop the next-generation technologies that change how billions of users connect, explore, and interact with information and one another. Our products need to handle information at massive scale, and extend well beyond web search. We're looking for engineers who bring fresh ideas from all areas, including information retrieval, distributed computing, large-scale system design, networking and data storage, security, artificial intelligence, natural language processing, UI design and mobile; the list goes on and is growing every day. As a software engineer, you will work on a specific project critical to Google’s needs with opportunities to switch teams and projects as you and our fast-paced business grow and evolve. We need our engineers to be versatile, display leadership qualities and be enthusiastic to take on new problems across the full-stack as we continue to push technology forward.

We are looking for a Staff Software Engineer to help our efforts in Recursive Self-Improvement (RSI), Automated Prompt Optimization (APO), and Agentic prompt simplification. We are working towards ensuring that multimodal journeys in Search are fully onboarded onto RSI, and we also motivate horizontal improvements across Search from our learnings.

In this role, you will help set and execute the technical direction as part of a strong team for autonomous, closed-loop agentic systems that continuously optimize, compress, and modularize system instructions and agent trajectories—improving instruction-following fidelity, reasoning quality, and inference efficiency at scale.

In Google Search, we're reimagining what it means to search for information – any way and anywhere. To do that, we need to solve complex engineering challenges and expand our infrastructure, while maintaining a universally accessible and useful experience that people around the world rely on. In joining the Search team, you'll have an opportunity to make an impact on billions of people globally.

Responsibilities

• Partner with senior leads to architect closed-loop RSI pipelines that autonomously diagnose failure modes, refine instructions/trajectories, synthesize targeted training/eval data, and verify gains without manual intervention.
• Investigate and build automated multi-agent systems that audit, compress, and de-conflict complex model context—maximizing token efficiency and inference speed while preserving behavioral and safety guardrails.
• Bridge inference-time context engineering with model post-training (SFT/RL) to distill complex rules into model weights.
• Advance automated optimization methods for production multimodal and reasoning systems.
• Help define the multi-quarter roadmap across Engineering and Research teams, establishing rigorous automated evaluation gates.

Minimum qualifications:

• Bachelor’s degree or equivalent practical experience.
• 8 years of experience in software development.
• 5 years of experience with ML design and with LLM/multimodal agent architectures.
• Experience utilizing agentic AI frameworks or developer kits to build and deploy autonomous AI systems.
• Experience in automated LLM evaluation (e.g., LLM-as-a-Judge, multi-turn user simulation, or Pareto optimization).
• Experience integrating generative AI tools or LLM interfaces into workflows and designing metrics for the AI tools.
  

Preferred qualifications:

• 8 years of experience with data structures and algorithms.
• 3 years of experience in a technical leadership role leading project teams and setting technical direction.
• 3 years of experience working in a complex, matrixed organization involving cross-functional, or cross-business projects.
• Experience building Recursive Self-Improvement, automated prompt, or agent refinement frameworks (e.g., DSPy, OPRO, TextGrad, Reflexion, STaR).
• Experience elevating performance on benchmarks for instruction-following, multimodal reasoning, and agents.
• Experience with prompt compression, structured output enforcement, or context-to-weight distillation.