Software Engineer (SWE) — LLM Superuser
We're recruiting an exceptional engineer to participate in a structured AI Uplift Trial — a research initiative measuring the productivity and quality gains that expert LLM usage confers on complex software engineering tasks.
This is not a role for someone who occasionally uses AI tools. We need an LLM Superuser: a practitioner with deep, intuitive command of how frontier models think, fail, and excel.
Engineering Baseline
4+ years professional software engineering experience.
Strong proficiency in at least two languages (Python, TypeScript, Go, Rust, etc.).
Comfortable across the stack: APIs, databases, cloud infrastructure, application logic.
LLM Expertise — This Is the Role
Expert command of prompt engineering: chain-of-thought, few-shot, ReAct, structured outputs, and beyond.
Experience building agentic systems: tool-calling loops, multi-step reasoning pipelines, orchestration frameworks.
Deep familiarity with context management: RAG, chunking, vector search, avoiding context degradation.
Ability to critically evaluate model outputs — catching hallucinations and subtle failures a casual user would miss.
Strong opinions about where LLMs genuinely help in engineering workflows and where they don't.
You can look at a hard engineering problem and immediately formulate an LLM strategy. You have a personal library of hard-won prompt patterns. You can explain why an approach works — and adapt it on the fly when it doesn't. You treat the model as a collaborator you understand deeply, not a black box you poke randomly.