AI Prompt Engineer

Description

The company is hiring an AI Prompt Engineer for a full-time, regular Information Technology role supporting U.S. Customs and Border Protection (CBP). No security clearance is needed to start.

The work applies generative AI to a border mission. The aim is to change how frontline personnel find information, make decisions and carry out their tasks, using AI tools and large language models to improve efficiency, situational awareness and access to information that matters to the mission.

You will sit in the company's Agile Solution Factory, a team that prototypes quickly, runs experiments and keeps learning. The ideal person turns hard operational problems into practical AI solutions that give mission users clear, measurable value.

Responsibilities

• Write and refine prompts for different generative AI models (LLMs) so they produce the behaviour, answers and creative output you need.
• Test prompt variations side by side, study the results and use them to keep improving.
• Work with several groups to build prompts into workflows that cut down manual work.
• Keep up with new developments in LLMs, generative AI and prompt engineering methods.
• Give input on, and help plan, model fine-tuning and Retrieval Augmented Generation (RAG) to raise accuracy.
• Diagnose unexpected model output and find the cause, whether it lies in the prompt structure, a model limit or the data context.

Requirements

• Required: You must be a U.S. citizen able to pass a CBP background investigation. The checks include, among others: no felony convictions in the past 3 years, no illegal drug use in the past 1 year, and no misconduct such as theft or fraud in the past 1 year.
• Required: You must be able to work onsite in Ashburn, VA, in Washington, D.C., or in both.
• Required: A bachelor's degree in computer science, software engineering, information technology or a related field. Equivalent experience may be accepted in place of the degree.
• Required: At least 1 year of software development work, an internship or relevant academic project work.
• Required: At least 1 year of hands-on work with large language models such as the OpenAI GPT series, Llama or Gemini, or with other generative AI models.
• Required: Working knowledge of at least one programming language, such as Python or JavaScript/TypeScript.
• Required: Familiarity with prompt engineering techniques such as few-shot prompting, chain-of-thought, structured outputs and function calling.
• Desired: Strong problem solving, with a data-driven approach to testing and optimisation.
• Desired: Clear written and spoken communication, including explaining complex technical ideas simply.
• Desired: Ability to work alone and with others in a fast-moving, changing environment.
• Desired: Exposure to RAG concepts, including embeddings and vector databases.
• Desired: Understanding of AI safety issues such as prompt injection, bias and data privacy.
• Desired: Exposure to Agile or Scrum development teams.

Conditions

The position depends on funding, so it may not be filled straight away. It reflects roles the company regularly has open, and applicants may be considered for other openings there.

Pay depends on factors including location, Federal Government contract labour categories and wage rates, prior relevant experience, specific skills, education and certifications.

Because the role can be based in more than one location, the posted range is the national average for the position.

Proposed salary range: $53,100-$106,300.

Travel: up to 10%, local only.

Benefits

• Flexible time off that you can use as you need it.
• Access to learning resources for career development.
• Health care, wellness, financial, retirement, family support, continuing education and time off benefits.