Senior Software Engineer (AI/LLM features) — Paid Research Study

This is expert-judgment work, not a coding contract. We're a small startup building developer tools for teams that ship AI-powered features, and we run formal research studies to validate our methods. For this study we need three engineers to independently review real-world AI prompt specifications using a written guide. What you'll actually do - Read a reference guide (provided as a PDF and in-app), then work through ~75 short classification items in our web app, each with a one-line justification quoting the text you relied on - Total commitment ≈ 5 hours including the guide, completed in 1–3 scheduled live sessions - Sessions are conducted on a video call with screen share — this is an integrity requirement of the study design, not supervision of your ability. Breaks between items are fine; recordings are retained privately as study records and never published Who we're looking for - A genuinely senior software engineer — senior/staff level and up, with a verifiable professional history. Not a manager who no longer touches code, and not QA/ops-only experience - You have personally built and shipped LLM-powered features in production — prompt design, structured outputs (JSON/schemas), tool use, RAG, agents, or similar. We will ask you to describe specific shipped work; links or concrete detail beat titles - Full professional English reading fluency. The entire task is close reading of dense English specifications; this is a hard requirement of the work itself - No AI/LLM-evaluation specialization required — if you ship AI features, you're the profile Terms (contractual, agreed in writing before starting) - You personally perform all work — no delegation, no assistance - No AI tools of any kind during the task (no ChatGPT/Claude/Copilot/etc.), affirmed verbally on each session recording - Time tracker on for all sessions - Payment is the full fixed price of $300 (≈ $60/hr) paid on completion regardless of your answers — we are studying the judgment process itself, and there are no "right answers" you're being scored against for payment purposes Selection process Short structured video interview (~15–20 min) including a brief live reasoning exercise. No preparation is possible or needed — that's the point of it. We're hiring three raters; selection is on profile fit and the interview. To apply, include in your proposal: 1. Two or three sentences on an LLM-powered feature you personally built and shipped — what it did, what your role was, and what was hard about it 2. Your seniority evidence: current title, years of experience, and anything verifiable (LinkedIn, GitHub, product links) 3. Your general availability over the next week for scheduled video sessions Applications without item 1 answered specifically will not be considered.