Staff+ Research Engineer – Recursive Self-Improvement Lead

STAFF+ RESEARCH ENGINEER — RECURSIVE SELF-IMPROVEMENT LEAD

About Us

Building the ranking intelligence layer for the internet.

The company builds Recursive Ranking Intelligence (RRI): an autonomous research engine in which a team of AI agents does the work of an ML research team, autonomously or together with ML researchers, in our clients' own clouds or on the company-hosted instances. The models RRI produces serve production traffic for the world's largest retailers, marketplaces and travel platforms, alongside the company's ranking platform, which runs frontier ranking models in production at sub-25ms latency and enterprise scale. Each gain compounds into revenue and margin lift measured in hundreds of millions of dollars per customer.

We are a small, highly technical, early-stage team turning recent advances in AI into production systems that operate under unforgiving real-world constraints.

About the Role

RRI already does ML research on real client problems, and the models it trains serve production traffic at some of the world's largest retailers.

We're looking for a Staff+ Research Engineer to join leadership of our work on recursive self-improvement (RSI): making RRI better at doing research, not only at doing more of it. You will set the research agenda, run the experiments, and ship what works into a system our clients run in production. This is an applied research role: you write the code yourself, and success is measured by what RRI delivers for clients' models. You will join our Recursive Self-Improvement team and help grow it as the area grows. You will work side by side with a Research Engineer focused on the RRI harness, often on the same projects; your profile adds depth in applied research.

Key Responsibilities

• Own the RSI agenda: Define the research roadmap for how RRI improves its own research process, choose the bets, and decide what ships.

• Advance agent strategy: Design and test exploration policies, compute allocation across parallel agents, and agents that revise their own instructions and strategy.

• Learn from past sessions: Turn recorded research sessions into training and evaluation signal.

• Define research quality: Specify what "better research" means and build the eval together with the rest of the RRI team.

• Compound knowledge: Make what RRI learns in one session improve the next, across tasks and clients.

• Drive research frontiers: Track the RSI, autonomous research, and agentic LLM literature, and turn promising ideas into measured experiments within weeks.

• Partner with scientists: Work closely with our applied scientists, who use RRI daily on real client problems, to find where the agents fall short.

About You

• Proven track record: Bring 7+ years of experience in applied ML research or research engineering, with hands-on work in at least one of LLM agents, AutoML, meta-learning, recursive self-improvement, or a closely related area.

• Research into production: Have taken research ideas from prototype to a measured improvement in a shipped product or production system.

• Experimental rigor: Think in baselines, ablations, seeds, and confidence intervals, and distrust any win you have not reproduced.

• Hands-on engineering: Write production-quality Python and build your own experiments end to end.

• Frontier LLM fluency: Understand how frontier models behave in long agentic loops — context limits, compaction, tool use, failure modes, and cost.

• Extreme ownership: Take absolute accountability for a research direction and navigate the ambiguity of an early-stage team.

• Technical leadership: Experience leading or mentoring a small research team.

Strong Candidates May Also Bring

• Research footprint: Peer-reviewed publications (e.g., NeurIPS, ICML, ICLR) or widely used open-source work in agents, RL, or AutoML.

• RL and LLM post-training: Experience with reinforcement learning or LLM post-training, such as RLHF, reward modelling, or training agents with RL.

• Evaluation expertise: Experience building evaluations or benchmarks for LLMs or agents, such as reward models, LLM-as-judge, or agentic task suites.

• Ranking domain exposure: Background in search, recommendation, or learning-to-rank models.

• Multi-agent systems: Experience with multi-agent systems or agent swarms, including their cost and safety trade-offs.

What We Value

• Rigorous scientific thinking: You measure before you claim, and you prefer a smaller, reproducible gain to a larger one inside the noise.

• Pragmatic speed: You move from paper to prototype to measured result quickly, without leaving a trail of unrepeatable experiments.

• Research that ships: You judge ideas by what they do for clients' models in production, not by novelty alone.

What We Offer

• High-impact influence: A staff-level role that shapes the direction of the company's autonomous research engine.

• Pioneering systems: The chance to do frontier RSI research on real production problems, with quantified client outcomes.

• Complete flexibility: Unlimited paid time off, flexible hybrid/remote configurations, and a highly collaborative, world-class engineering culture.