DeepLLMData specializes in powering frontier machine learning and AI models with high-quality expert human data. Our offerings include expertise in Structured Fine-Tuning (SFT) and Reinforcement Learning with Human Feedback (RLHF), as well as support for STEM, Domains, mathematics, and various industry-specific areas.
Role Description
We are seeking experienced US Legal Professionals / Paralegals to design, develop, and evaluate high-quality prompts for Large Language Models (LLMs) using real-world legal documents and frameworks. This role combines legal expertise with AI prompt engineering and response validation.
Key Project Expectation
- Create legal-domain prompts using structured and unstructured legal content such as:
Contracts and agreements (NDAs, MSAs, SLAs, employment contracts) Case law and legal precedents
Statutes, regulations, and compliance documents
Legal opinions, briefs, and memos
- Design prompts that simulate real-world legal scenarios, including:
Contract review and risk identification
Clause interpretation and drafting
Legal research and case summarization
Compliance analysis
Regulatory impact assessments
- Test and validate prompts across leading LLMs such as:
- Evaluate model responses for:
Legal accuracy and correctness Logical reasoning and argument strength
Alignment with applicable laws and jurisdictions
Risk identification and compliance considerations
- Develop “Golden Responses” (ideal outputs) that serve as benchmarks for model evaluation and training.
Required Skills & Qualifications
- 2–10 years of experience in US Legal / Paralegal / Compliance roles
- Strong experience with: Contract drafting and review
Legal research and documentation
Regulatory and compliance frameworks
- Robust analytical and critical thinking skills to evaluate AI-generated legal outputs