The Vision: System-Level Optimization

Most current RSI work is highly LLM-centric, treating model weights as the sole unit of improvement. Every step inherits the massive cost of a training run, and compounding progress arrives late. We take a whole-system view: the LLM is just one component of a larger reasoning system that includes code, prompts, search strategies, and tool use. We are building a self-optimizing optimizer—a system where every task it tackles supplies the signal needed to optimize its own orchestration code.

As an AI Engineer, you will build a robust, scalable platform for our intelligence on top of foundation models. You will architect a complete system that surrounds the LLMs, managing the complex, multi-step interactions required to extract and synthesize information. You must be comfortable working in, and building, the high-performance infrastructure that makes our fast, self-improving reasoning possible.

You are a good fit if you:

• Think and breathe Python. You are a Python engineer proficient in building complex, agentic systems or multi-step, stateful execution frameworks.

• Work autonomously to develop both clearly defined and ambiguous ideas, including your own, into reality.

• Excel at designing and building reliable, high-performance infrastructure that interacts heavily with external, third-party LLMs – some experimental, some large-scale and publicly deployed.

• Can architect clean abstractions for complex workflows, specifically synthesizing fragmented information gathered over thousands of parallel, asynchronous queries.

• Care deeply about code quality, performance profiling, and building the stable, scalable platform that allows research to run autonomously.

Our Results & Impact

Our system-level RSI reaches state-of-the-art (SOTA) performance without modifying a single LLM parameter:

• 12.3% boost on frontier models for LiveCodeBench Pro, setting a new SOTA at 93.9%, among many other SOTA results we have shared on our blog at .

• Universal improvement: Every model tested improved, proving our harness encodes highly transferable task structure.

• Dominated 6 unseen benchmarks spanning competition mathematics, scientific coding, long-horizon planning, agentic tool use, and long-context retrieval—all automatically.

Culture of Transparent, Inspectable AI

We prioritize explainability. By running our optimization loops at the system level rather than baking them into uninterpretable parameters, every improvement remains human-readable—transparent code, explicit prompts, and clear data. We believe fast, powerful RSI is fully compatible with tighter, more deliberate oversight. This approach values thoughtful, rigorous diagnostic engineering over blindly scaling compute and black-box models.

Our Team & Engineering Culture

We are a high-leverage team of 10 engineers and researchers. We thrive in an in-office environment built around high-bandwidth collaboration, rapid whiteboarding, and low-ego problem solving.

Our engineering culture is highly collaborative, mentorship-driven, and deeply inclusive. We value clear communication, rigorous testing, and deliberate architectural design. At Poetiq, you won't just be optimizing weights on the periphery; you will be core to designing the interpretable reasoning architectures of the future.