Prompt Engineer

About the role

Cantina Labs is at the forefront of social AI, creating innovative models that enhance storytelling and connection through AI characters. As a Prompt Engineer, you will play a vital role in shaping user experiences by collaborating with various teams to develop engaging and realistic AI interactions.

Key facts

What you'll do

• Create and refine prompts that dictate how Cantina's AI characters convey personality, emotions, and social interactions across various media formats, including text, voice, and video.
• Determine essential personality traits that make AI characters appealing and develop datasets and prompt frameworks to consistently replicate these traits.
• Experiment with new character traits and features through prompt engineering, using both qualitative and quantitative methods for evaluation.
• Balance immediate product development needs with the long-term goals of behavioral consistency and quality within the Cantina ecosystem.
• Collaborate with research teams to establish and maintain evaluations that track bot performance, personality accuracy, and interaction quality over time.
• Work alongside product, design, and engineering teams to translate creative concepts and user feedback into scalable prompt systems.
• Review and enhance prompts as they transition to production, contributing to the establishment of best practices and launch criteria.
• Monitor and address emerging issues related to bot behavior, prioritizing improvements and implementing prompt-level solutions in complex or high-impact situations.
• Document insights on model behaviors, prompting strategies, and guidelines to facilitate prompt development across different teams.
• Assist research and machine learning efforts by providing meta-prompts for synthetic data generation and evaluating new model behaviors.

Requirements

• A minimum of 3 years of experience in prompt engineering and developing features based on large language models (LLMs) from start to finish.
• Proven experience in writing, testing, and refining complex multi-step and system-level prompts.
• In-depth knowledge of LLM capabilities, limitations, and nuances, with a strong understanding of how input design influences model behavior.
• Experience in creating datasets and assessing model outputs in partnership with machine learning or product teams.
• A solid foundation in psychology, sociology, linguistics, or related disciplines relevant to personality, behavior, and dialect.
• A systematic, hypothesis-driven approach to experimentation and iteration.
• Ability to juggle multiple projects, prioritize tasks effectively, and work independently in fast-paced, ambiguous environments.
• Strong communication skills, capable of articulating prompt design choices to both technical and non-technical audiences.
• Comfort in engaging with executive-level stakeholders.
• Basic familiarity with Python and the ability to navigate unfamiliar codebases.

Compensation

The expected annual base salary for this position ranges from $150,000 to $180,000. Several factors will influence the final compensation, including skills, experience, job responsibilities, location, and prevailing market rates.

Benefits

• Competitive salary with substantial equity options
• Comprehensive medical, dental, and vision insurance, with nearly all premiums covered by Cantina
• 42 days of paid time off, consisting of:
• 15 personal days
• 10 sick days
• 15 company holidays
• 2 floating holidays
• Generous parental leave and fertility assistance
• 401(k) retirement savings plan
• Lifestyle spending account offering $500 per month for personal use
• Complimentary lunch and snacks for employees in the office
• One Medical membership, and additional benefits!