AI Engineer (x2) – Legal Innovation Team
Application Deadline: 29 May 2026
Department: Information Technology
Employment Type: Full Time
Location: 10 Earlsfort Terrace, Dublin 2, D02 T380
Description
The AI Engineer is the technical builder within the Legal Innovation Team. The work is fundamentally applied natural language processing (NLP): extracting meaning from legal documents, classifying and comparing text, reasoning over complex material, and building systems that analyse language at scale.
They take the specifications and requirements defined by Innovation Lawyers and turn them into working solutions: document processing pipelines, AI‑powered analysis workflows and deployed systems that lawyers and clients use daily.
This is not a generic software engineering role. The AI Engineer works exclusively on legal technology, embedded in a team that understands the domain deeply. They do not need to be a lawyer (that expertise sits with the Innovation Lawyers they work alongside) but they must build systems that meet the exacting standards legal work demands: accuracy, reliability, auditability, and data security.
Key Responsibilities
Building document processing pipelines. Extracting text from PDFs, parsing structured and unstructured legal documents, transforming raw content into clean, queryable data. This is foundational work: every AI‑powered solution starts with getting the data right.
Designing and building NLP solutions for legal documents. Summarisation, classification, comparison, risk analysis, and information extraction, using large language models and other NLP techniques as appropriate. The output is always structured, auditable, and fit for use in legal work.
Working with LLM APIs. Selecting appropriate models, designing effective prompts and processing chains, parsing structured outputs, implementing guardrails, and optimising for cost and performance. LLMs are the primary tool, but not the only one; the right approach depends on the task.
Building web applications and user interfaces. Creating tools that lawyers and clients interact with directly, backed by databases, authentication, and audit logging.
Designing evaluation frameworks. Building test datasets, defining quality benchmarks, and systematically measuring AI output accuracy rather than relying on spot‑checking.
Database design and data modelling. Structuring data stores that support both the deterministic extraction stages and the AI analysis stages of a pipeline, with clear separation of concerns and full auditability.
Building internal tools that replace or improve upon expensive external services, with a focus on practical solutions that non‑technical users can operate confidently.
Integrating solutions with the firm’s technology stack, connecting AI capabilities with document management systems, Microsoft 365, and client‑facing platforms.
Ensuring security and compliance. All solutions must meet the firm’s data protection, confidentiality, and information security requirements. Legal data is among the most sensitive there is.
Maintaining and iterating on deployed solutions based on user feedback and evolving requirements.
About the Person
Essential
Strong proficiency in Python. The primary language our solutions are built in, used for everything from data pipelines to web applications to LLM integration.
Demonstrated experience with applied NLP. Text classification, information extraction, summarisation, document comparison, or similar tasks. NLP is the core discipline of this role; large language models are the primary tool, but the engineer must understand the broader landscape of techniques and know when simpler approaches outperform generative models.
Hands‑on experience working with large language models. Prompt engineering, API integration, structured output parsing, and systematic evaluation of model performance. We are looking for someone who has built real projects with LLMs, not just experimented casually.
A solid understanding of transformer architecture and how modern language models work. Not because we train models, but because understanding the underlying mechanics leads to better system design, more effective prompting, and a realistic sense of what LLMs can and cannot do reliably.
Experience with data processing and pipeline design. The ability to take messy, real‑world data (PDFs, scanned documents, inconsistent formats) and build robust pipelines that produce clean, structured output. Comfort with SQL and data manipulation libraries is expected.
Understanding of RAG architectures. Document chunking strategies, embedding models, vector databases, retrieval strategies, and the practical challenges of making retrieval‑augmented generation work reliably on real‑world documents.
Familiarity with evaluation and benchmarking methodology. Designing test sets, measuring AI system performance rigorously, and building feedback loops that catch regressions. The nature of legal work means AI outputs must be held to a high standard, and the engineer must be able to prove they meet it.
Version control and collaborative development. Proficiency with Git and experience working in a shared codebase.
A collaborative working style. This role is defined by working closely with non‑technical colleagues (Innovation Lawyers). The ability to listen to requirements, ask the right questions, and translate domain knowledge into technical design is essential.
Desirable
A postgraduate qualification in computer science, NLP, machine learning, or a related field, or equivalent depth of expertise demonstrated through professional work and projects.
Experience building document processing pipelines. PDF parsing, OCR, text extraction, and structured data extraction from unstructured documents.
Experience with a range of NLP techniques beyond LLM prompting, such as classification pipelines, similarity measures, or heuristic‑based extraction, and the judgement to choose the right tool for the task.
Exposure to agent frameworks and multi‑step AI workflows (e.g. LangChain, LangGraph, Claude tool use, OpenAI function calling).
Experience with data visualisation and building dashboards or reporting interfaces that make complex data accessible to non‑technical stakeholders.
Familiarity with cloud platforms, particularly Microsoft Azure, given the firm’s infrastructure. Experience with Azure OpenAI Service or equivalent cloud AI services is a plus.
A published research contribution (paper, open‑source project, or technical blog) demonstrating depth of engagement with AI/ML topics.
Experience communicating technical findings to non‑technical audiences, through reports, presentations, or stakeholder briefings.
Mindset
Solving real problems for real users. An engineer who cares about practical outcomes, not just building technically impressive systems. The best solution is often the simplest one that works reliably.
Comfortable working in an unfamiliar domain. Legal work is complex and nuanced, and the right person is curious about it rather than intimidated by it.
Balancing rigour with pace. Legal technology must be reliable enough to trust with client work, but the speed of AI development means the team cannot afford to over‑engineer every solution.
Self‑directed and resourceful. This is a small team building new things. There is no established playbook. The right person thrives in that environment.
A natural communicator who can explain technical concepts clearly to lawyers and stakeholders who are not engineers.
Why This Role
High‑impact, visible work. Every solution you build will be used by lawyers on real matters with real clients. The feedback loop is short and the impact is tangible.
Cutting‑edge technology. Legal AI is one of the most active applied AI domains. You will work with the latest models and techniques on genuinely challenging problems: long documents, complex reasoning, high accuracy requirements, and real‑world constraints.
Domain depth over breadth. Rather than building generic AI applications, you develop deep expertise in a specific, intellectually rich domain. Legal reasoning, contract analysis, and regulatory compliance are hard problems that reward sustained focus.
Autonomy and ownership. In a small, specialised team, you own your solutions end‑to‑end, from design through deployment. There is no handoff to another team.
A stable, well‑resourced environment. Arthur Cox is one of Ireland’s leading law firms. This is not a startup that may not exist in 12 months. It is a long‑established institution investing seriously in AI, with the resources and client base to support ambitious work.
Reporting and Structure
The AI Engineer reports to the Senior Innovation Lawyer and works as part of the Legal Innovation Team.
Compensation & Benefits
Competitive salary commensurate with experience, benchmarked against AI/ML engineering roles in the Dublin market.
The firm offers a comprehensive benefits package.
#J-18808-Ljbffr