Role Overview

Key Responsibilities

• Design, implement, and refine prompts, system messages, and tool instructions for customer‑facing GenAI features.
• Decompose complex user intents into modular prompt chains, deciding when to call tools, retrieve knowledge, or delegate to agents.
• Own prompt evaluation by building datasets, tagging feedback, and writing custom evals. Use a mix of automation and human review to improve quality fast.
• Run adversarial and safety testing sessions to identify weaknesses, then iterate to improve robustness, tone, and compliance.
• Collaborate daily with product managers, designers, and AI platform engineers to ship experiments fast and learn from data.
• Stay current with new models, frameworks, and research, sharing insights with the wider AI team.

Qualifications

• 2+ years of experience writing production code in JavaScript, TypeScript, or Python.
• 1+ year hands‑on experience with large language models and prompt engineering.
• Experience combing through user feedback, building evaluation datasets, and designing practical evals for non-deterministic systems. Skilled at debugging unpredictable model behavior in production
• Proven experience working with multiple LLMs and AI models
• Strong written English and conversation‑design skills, able to craft clear, product‑aligned instructions.
• Excellent collaboration and communication skills, with a track record of working effectively in cross-functional teams.
• Product mindset with an iterative approach to experimentation and measurement.
• Experience with RAG pipelines and their integration into AI‑driven applications, including search and personalization workflows.
• Passion for staying at the forefront of AI infrastructure and applying new technologies to solve real-world problems at scale.

Preferred Qualifications

• Background in linguistics, UX writing, or conversation design.
• Knowledge of safety frameworks, privacy constraints, and compliance requirements.
• Prior work on consumer productivity tools or other SaaS products.
• Adversarial prompting or safety red‑teaming experience.