Legal AI Specialist (Legal AI Platform)

About Us
Datica is an AI-powered LegalTech platform that leverages Large Language Models (LLMs), AI Agents, and Retrieval-Augmented Generation (RAG) to transform legal workflows.

Our platform is designed to help individuals, lawyers, legal teams, and businesses conduct legal research, understand legal issues, analyze documents, draft legal content, and complete complex legal tasks more efficiently.

We are looking for a hands-on Legal AI Specialist to help ensure that Datica’s outputs are legally sound, reliable, well-reasoned, and useful in real-world legal work.

This is not a traditional legal counsel position. The Legal AI Specialist will work at the intersection of law, artificial intelligence, product development, and legal knowledge management.

The successful candidate will improve and develop system prompts, prepare legal datasets, design AI-assisted legal workflows, evaluate model responses, and work closely with the product, AI, and engineering teams to improve Datica’s legal capabilities.

Responsibilities
Legal Prompt Development

• Review, test, and improve Datica’s existing system prompts.
• Develop new system prompts for different legal domains, user groups, and legal tasks.
• Translate legal reasoning processes, professional standards, and practical legal knowledge into clear instructions for AI systems.
• Define the expected structure, reasoning approach, tone, limitations, and output format for different legal tasks.
• Develop prompt strategies that help the system identify relevant facts, ask appropriate follow-up questions, apply applicable laws, distinguish uncertainty, and produce structured legal analysis.
• Create domain-specific prompts for areas such as contracts, corporate law, commercial disputes, employment law, real estate, civil procedure, criminal law, family law, and other relevant fields.
• Test prompts across different factual scenarios, user formulations, document types, and edge cases.
• Diagnose weaknesses in AI responses and determine whether they result from prompts, retrieved sources, missing context, workflow design, or model limitations.
• Maintain prompt versions, testing records, evaluation results, and documented changes.

Legal Quality Assurance and Evaluation

• Review and score Datica’s responses for legal accuracy, completeness, reasoning quality, relevance, clarity, and practical usefulness.
• Verify whether legal conclusions are properly supported by applicable statutes, regulations, judicial decisions, and other authoritative legal sources.
• Identify hallucinations, unsupported claims, outdated legal rules, incorrect citations, missing exceptions, and flawed legal reasoning.
• Develop structured evaluation rubrics for different legal tasks and user groups.
• Create benchmark questions, model answers, scoring criteria, and failure classifications.
• Compare responses produced by different prompts, models, RAG configurations, and agent workflows.
• Conduct blind and comparative evaluations of AI-generated legal responses.
• Design challenging test cases, ambiguous scenarios, adversarial questions, and edge cases.
• Maintain regression test sets to ensure that product updates do not reduce legal quality.
• Classify recurring errors and work with AI and engineering teams to address their root causes.
• Define minimum quality thresholds for the release of new legal features and workflows.
• Participate in periodic quality reviews of Datica’s production outputs.

Legal Dataset Development

• Identify, collect, organize, clean, classify, and structure legal materials for use in RAG, evaluation, model improvement, and product development.
• Prepare high-quality datasets from statutes, regulations, judicial decisions, legal opinions, contracts, pleadings, legal forms, commentaries, and other relevant sources.
• Develop taxonomies, metadata structures, labels, and annotation guidelines for legal documents and legal tasks.
• Create question-and-answer datasets based on Iranian law and realistic legal scenarios.
• Draft reference answers and legal reasoning explanations for benchmark and evaluation datasets.
• Review and validate legal annotations prepared by other team members or external contributors.
• Identify conflicts, amendments, repealed provisions, duplicate documents, and inconsistencies within legal data.
• Help define the relationships among laws, regulations, judicial decisions, legal concepts, procedures, and document types.
• Ensure datasets are legally accurate, properly sourced, representative of real use cases, and suitable for their intended AI application.
• Contribute to the anonymization and safe handling of judgments, case files, contracts, and other potentially sensitive documents.
• Help develop quality-control processes for large-scale legal data collection and annotation.

Legal Workflow and Product Design

• Study how lawyers, legal departments, businesses, and individuals currently perform legal tasks.
• Convert complex legal processes into structured, AI-assisted workflows.
• Design workflows for legal research, legal consultation, contract drafting, contract review, document analysis, litigation preparation, compliance, legal due diligence, and other professional tasks.
• Break down legal work into appropriate stages, decisions, inputs, outputs, review points, and escalation rules.
• Determine what information the system should collect from the user before generating an answer or document.
• Define when Datica should ask follow-up questions, retrieve additional sources, use a specialized agent, produce a warning, or require human review.
• Design legal templates, checklists, questionnaires, decision trees, and structured outputs.
• Help distinguish workflows suitable for the general public from those designed for lawyers, businesses, and professional legal teams.
• Work closely with the Product Team to ensure legal workflows are understandable and intuitive.
• Work with the Product Team to prioritize features based on user needs, legal value, feasibility, and risk.
• Participate in interviews and usability tests with lawyers, legal teams, corporate users, and other target customers.
• Translate user feedback into concrete improvements to prompts, datasets, workflows, and product features.

Cross-Functional Collaboration

• Work closely with AI engineers on prompt behavior, RAG performance, agent logic, evaluation systems, and model limitations.
• Collaborate with backend and frontend engineers to ensure legal requirements are accurately implemented.
• Work with the Product Team to create trustworthy and effective legal user experiences.
• Collaborate with the Product Team on product strategy, quality standards, and the legal product roadmap.
• Explain legal concepts and requirements clearly to non-lawyers and technical team members.
• Review proposed features to identify legal-quality, user-safety, privacy, and professional-responsibility risks.
• Help establish internal standards for developing, testing, approving, and monitoring AI-powered legal features.
• Contribute to documentation and training materials for internal teams and users.

Requirements

• LL.B. or an equivalent degree in law from a reputable university.
• At least 3 years of meaningful legal experience in a law firm, corporate legal department, judiciary-related role, legal research position, or another relevant professional environment.
• Strong knowledge of Iranian law and the Iranian legal system.
• Strong legal research, legal writing, analytical, and issue-spotting skills.
• Experience using ChatGPT, Claude, Gemini, or other LLM products for professional legal work.
• Ability to analyze complex facts and produce clear, structured, and well-supported legal reasoning.
• Experience reviewing contracts, legal opinions, pleadings, legal memoranda, or other professional legal documents.
• High attention to detail and a strong commitment to legal accuracy.
• Ability to identify missing facts, exceptions, conflicting rules, procedural considerations, and alternative interpretations.
• Ability to explain complex legal matters clearly to both legal and non-legal audiences.
• Comfort working with structured data, spreadsheets, annotation tools, and repetitive testing processes.
• Genuine interest in artificial intelligence, LegalTech, and the transformation of legal work.
• Ability to learn how LLMs, prompts, RAG systems, and AI agents operate at a practical level.
• Ability to work collaboratively with product managers, designers, software engineers, and AI engineers.
• Ability to work independently, take ownership, and manage multiple legal domains and workstreams.
• Strong written and verbal communication skills in Persian.
• Good working proficiency in English, particularly for reading technical documentation and international LegalTech materials.

Nice to Have

• Master’s degree in law or specialization in a relevant legal field.
• Iranian Bar membership or meaningful professional legal-practice experience.
• Experience with legal annotation, taxonomy development, document classification, or knowledge-management projects.
• Familiarity with prompt engineering and the practical behavior of LLMs.
• Familiarity with RAG, embeddings, vector search, AI agents, or model evaluation.
• Familiarity with evaluation concepts such as benchmark datasets, golden answers, grading rubrics, inter-rater agreement, regression testing, and error taxonomies.
• Experience working in a startup, technology company, or interdisciplinary product team.
• Experience designing internal legal processes, standard operating procedures, contract playbooks, or legal-service workflows.

Prior technical or programming experience is not required. However, the candidate must be willing and able to understand the basic operation of LLMs, RAG systems, prompts, AI agents, and evaluation pipelines.

What Success Looks Like
During the first few months, the Legal AI Specialist will be expected to:

• Develop a deep understanding of Datica’s existing prompts, legal sources, target users, and core product workflows.
• Audit the existing system prompts and identify the most important legal-quality weaknesses.
• Establish a consistent methodology for testing, scoring, and improving legal responses.
• Develop clear evaluation rubrics covering legal accuracy, source support, reasoning, completeness, relevance, clarity, and user safety.
• Create an initial benchmark suite representing Datica’s most important legal use cases.
• Improve the quality and consistency of responses in priority legal domains.
• Establish a documented prompt-development and version-control process.
• Prepare high-quality datasets and reference answers for priority legal tasks.
• Identify recurring failure patterns and work with the AI team to reduce them.
• Design at least several end-to-end legal workflows for lawyers and corporate legal teams.
• Establish a repeatable legal quality-assurance process for new product releases.

What We Offer

• Work on a cutting-edge AI and LegalTech product.
• Opportunity to shape how artificial intelligence is applied to Iranian law.
• Meaningful ownership over Datica’s legal intelligence, evaluation standards, and knowledge systems.
• Opportunity to develop new approaches to legal research, legal drafting, and professional legal workflows.
• Direct collaboration with the founder, AI engineers, product designers, and software engineers.
• Opportunity to shape the product from MVP to scale.
• Collaborative, interdisciplinary, and product-driven culture.
• Flexible working hours.
• Significant learning opportunities at the intersection of law, AI, product development, and data.