Job DescriptionWe are seeking a Prompt Engineer with a focus on Trust & Safety to join our growing AI team. In this role, you will design and evaluate LLM-based systems to automate human review processes, filter harmful content, and enhance the quality and reliability of model responses. You'll collaborate directly with client teams, product owners, and engineers to develop safe, scalable, and contextually aligned AI solutions. Your work will directly influence how large language models are deployed responsibly across high-impact domains including search, content generation, education, and customer service.
As a key technical subject matter expert, you will work hands-on to design and test prompts, automate workflows, and drive improvements in the accuracy and reliability of AI-driven solutions. This is a collaborative role with a focus on tailoring solutions to meet client needs, helping them scale while maintaining high standards for quality.
Responsibilities
- Design and iterate on prompts that enable LLMs to:
- Translate content while flagging idiomatic or culturally sensitive expressions
- Classify text by tone, sentiment, or formality
- Filter or suppress unsafe, biased, or harmful outputs (e.g., nudity, violence, hate speech)
- Implement few-shot and one-shot learning strategies to improve model behavior in edge cases
- Red team LLMs to uncover safety vulnerabilities and help build more robust response boundaries
- Develop and evaluate safety-centric prompt pipelines using tools like LangChain, OpenAI API, and Hugging Face
- Collaborate with internal stakeholders and external clients on prompt testing strategies and annotation workflows
- Analyze output performance using structured evaluation metrics (e.g., precision, safety scores, coverage)Lead or contribute to the development of content moderation pipelines using LLMs
- Apply and enforce safety guidelines and policy-based reasoning in prompt design
- Design LLM personas that reduce bias and maintain tone control
- Work with internal teams to red team models, ensuring compliance with organizational safety and fairness standards
- Build tools that support automated content flagging, labeling, and escalation
Technical Skills & Qualifications:
- 2+ years experience in prompt engineering, LLM fine-tuning, or NLP/ML evaluation
- Proficiency with Python and tools such as OpenAI API, Hugging Face, LangChain, LlamaIndex
- Experience designing prompts for classification, translation, and behavioral alignment
- Strong understanding of few-shot, one-shot, and zero-shot learning techniques
- Familiarity with Trust & Safety domains: content moderation, harmful content detection, policy compliance, red teaming
- Experience working with JSON data, evaluation pipelines, or annotation workflows
- Excellent written and verbal communication skills; able to explain prompt design decisions clearly
About You:
- You are comfortable jumping into client conversations and meeting them where they are, bringing technical expertise to guide decisions and iterate on solutions.
- You have experience designing and implementing solutions using LLMs, specifically with a focus on automating processes and improving workflows.
- You are proactive, hands-on, and ready to contribute both technically and strategically to improve our client's systems.
We're an equal opportunity employer committed to increasing diversity and inclusion in today's workforce. All qualified applicants will receive consideration for employment without regard to race, color, religion, gender, gender identity or expression, sexual orientation, national origin, genetics, disability, age, or veteran status. Minorities, women, LGBTQ candidates, and individuals with disabilities are encouraged to apply. If you require an accommodation, please review our accessibility policy and reach out to our accessibility officer with any questions.
