Manager, Prompt Engineering

At Scale AI, we accelerate AI development by working at the forefront of AI, trust, and safety. We partner with leading businesses like OpenAI, Lyft, Meta, and more to develop LLM applications, improve training datasets, and make models more trustworthy and reliable.

Role

We are seeking a Prompt Engineering Manager to lead a team of experimental engineers who focus on application and use of AI in novel, creative ways. You will build skunkworks projects using state-of-the-art AI to support operations, research, and model evaluation teams. As prompt engineering evolves, you'll continuously learn and master the latest models, explore AI behaviors, and apply creativity in your work. You'll elevate your team's talent, collaborate across teams, and drive innovation in AI applications.

You will:

• Lead and manage the Prompt Engineering team, fostering innovation and experimentation.
• Develop LLM applications to solve complex problems and improve data quality.
• Identify and address vulnerabilities in LLMs to enhance robustness.
• Document model behaviors, including vulnerability and adversarial testing.
• Collaborate with cross-functional teams to drive projects from concept to deployment.
• Mentor team members, enhancing their skills in prompt engineering.
• Analyze and communicate experimental results to stakeholders.

Ideally you’d have:

• 5+ years of engineering experience.
• Hands-on experience with state-of-the-art AI models.
• Proficiency in Python, TypeScript, or JavaScript.
• Experience building LLM-based applications or leveraging LLMs creatively. Hobby projects are ok, the more creative in their use of AI the better.
• Leadership skills with some management or mentoring experience.
• Strong communication and collaboration abilities.
• A creative mindset with a passion for experimentation.

Nice to haves:

• Experience with AI/LLM development or red teaming.
• Familiarity with LLM evaluation metrics and statistical analysis.
• Academic work related to language models or adversarial robustness.
• Ideally you'd be able to collaborate in person with our teams in SF or NYC at least once a month