Prompt Engineer AI
San Francisco, CA; New York, NY
You will: Create LLM apps to assess the quality of data, work through challenging issues involving synthetic data capabilities, carry out business solutions using "Jailbreak," and identify weaknesses in LLMs and other generative AI models. Create training materials for adversarial analysis and red teaming, and instruct operating staff in them.
Create Python/TypeScript scripts for data creation and vulnerability testing; collaborate with software engineers and ML engineers for reliable implementation and in-platform deployment.
Examine and display the outcomes of the experimental prompt.
Create and instruct software developers and ML engineers on new prompt engineering techniques.
You possess: a Bachelor's degree in mathematics or a language model-related field (linguistics, philosophy, etc.)
Ability to write TypeScript or Python scripts for APIs
Capacity to use statistics and data visualization to analyze model performance
familiarity with applying programmatically frontier-scale LLMs (GPT, Claude, PaLM) to particular applications or implementations
demonstrated aptitude for creating LLM-based applications or for breaking into LLMs to accomplish unique goals
knowledge of LLM evaluation statistics and metrics (BLUE, ROUGE, MAUVE, etc.)
strong communication abilities both in writing and speaking
Using creativity to spur internal research and development and experimenting
Desirables: Prior expertise with red teaming or AI/LLM developmental work is very advantageous.
Academic research on adversarial robustness in language models
Earlier reports of security flaws or weaknesses in LLMs or other AI models
in creating machine learning models in a cloud environment and utilizing cloud technology stacks, such as AWS or GCP
Salary - $140,00 - $210,000
