Director Artificial Intelligence Testing Jobs (NOW HIRING)

Founded in 1920, Akerman is recognized as one of the nation’s premier law firms, with more than 700 lawyers across the United States. 

Akerman LLP,  an AmLaw 100 firm, is seeking a hands-on Director of Artificial Intelligence to lead the design, development, and responsible deployment of AI systems across the firm. This is not a purely strategic or advisory position—we are looking for a builder. The successful candidate will have personally architected and shipped production AI applications, including agentic systems, can read and write code, and understands the practical realities of running AI on confidential, privileged data in a regulated professional-services environment.

The Director will own the process of scoping designing agentic workflows and tools that retrieve, reason over, and act on information across heterogeneous sources, selecting and tuning the right models for each task, and doing so under rigorous security and compliance controls. This person will work shoulder-to-shoulder with attorneys, practice groups, knowledge management, security, and IT.

Core Responsibilities:

Design and ship agentic AI systems. Architect, build, and operate agentic AI applications; systems that plan, call tools, retrieve and act on information, and execute multi-step workflows with appropriate human oversight. Build and maintain the orchestration layer (tool/function calling, multi-agent coordination, memory, state management, retries, and guardrails), and integrate agents with firm systems via MCP (Model Context Protocol) servers and other tool interfaces. Define where agents operate autonomously versus where a human stays in the loop.

Build the data and tooling backbone. Develop production-grade pipelines and tools that collect and process information from diverse sources including public and subscription websites, REST and streaming API feeds, MCP servers and feeds, email systems, document management systems, and SQL and vector databases. Own these systems end to end, including retrieval (RAG) architectures, evaluation, observability, and iteration.

Model selection and orchestration. Demonstrate working fluency with frontier foundation models (e.g., OpenAI, Anthropic Claude, Google) via API, as well as locally hosted open-weight models (e.g., Llama, Mistral, Qwen). Make sound, cost-aware decisions about which model and which agent design fit each use case, and route tasks accordingly.

Tune and operate open-weight models. Hands-on experience fine-tuning, adapting (LoRA/PEFT), quantizing, and serving open-weight models on firm-controlled or private-cloud infrastructure to meet specific practice and business needs—particularly for agentic tasks and where data sensitivity precludes sending information to third-party APIs.

Protect privilege and prevent data leakage. Treat the protection of attorney-client privilege, work product, and confidential client information as a first-order design constraint, made more acute by agentic systems that take actions and traverse multiple data sources. Architect agents and pipelines to prevent data leakage to external model providers, constrain tool permissions and scope, avoid inadvertent waiver or spoliation of privilege, enforce data residency and retention requirements, and maintain clear audit trails of every agent action.

Security partnership. Work closely with the firm's Information Security, Research (KM) and IT teams to ensure all AI systems, especially autonomous agents with tool access meet the firm's security standards, client outside-counsel guidelines, and audit requirements. Conduct or support AI risk assessments and threat modeling (including prompt injection, tool-abuse, excessive-agency, data-exfiltration, and model-supply-chain risks), and lead vendor security reviews.

Governance and compliance. Work with the firm's Information Security team that is responsible for  firm's AI governance framework, to keep the firm aligned with recognized standards for AI management systems (e.g., ISO/IEC 42001) and applicable regulatory regimes.

Cross-functional collaboration and enablement. Partner with firm management, practice groups and KM to identify high-value use cases, translate legal workflows into agentic designs and technical requirements, and prioritize by ROI, feasibility, and risk. Design AI applications training for attorneys and staff, lead change management, and build a culture of responsible, confident AI adoption.

Vendor and platform evaluation. Evaluate legal-AI vendors and agent platforms, distinguish substance from marketing, run pilots with measurable success criteria, and advise on build-vs-buy decisions.

Team leadership. Build, mentor, and manage a small team of engineers and/or AI specialists as the function grows.

Required Qualifications:

• Minimum 5 years of technology experience, including a minimum 3 years building and deploying production AI/ML applications. This must be practical, hands-on experience, not solely academic or research background. Candidates should be prepared to discuss systems they have personally built.
• Demonstrated experience building agentic AI systems. Tool/function calling, multi-step or multi-agent orchestration, and integrating agents with external systems and data sources (including MCP).
• Strong software engineering skills, including Python, working with REST/streaming APIs, SQL and vector databases, and modern AI/LLM and agent frameworks.
• Hands-on experience with RAG pipelines, embeddings, evaluation, and observability for LLM and agent applications.
• Working knowledge of  utilizing open-weight models, including on-premises or private-cloud GPU infrastructure.
• Demonstrated experience designing AI systems under strict security, privacy, and confidentiality constraints; familiarity with data-leakage prevention, least-privilege tool access, encryption, and audit logging.
• Working knowledge of AI governance and risk frameworks (e.g., ISO/IEC 42001, NIST AI RMF) and relevant data-privacy regulation.
• Proven ability to collaborate effectively with security, IT, and non-technical stakeholders, and to communicate technical concepts to attorneys and firm leadership.

Preferred Qualifications:

• Prior experience in a law firm, legal-technology provider, or other regulated professional-services or highly confidential environment.
• Understanding of attorney-client privilege, work-product doctrine, and the legal and ethical duties (e.g., ABA Model Rules and recent formal opinions on generative AI) that constrain how AI may be used in legal practice.
• Experience building MCP servers and designing evaluation/guardrail frameworks for autonomous agents.
• Familiarity with legal-specific platforms and use cases (document review, contract analysis, legal research, drafting).
• Background that combines software/AI engineering with exposure to litigation, eDiscovery, or knowledge management.
• Degree in computer science, engineering, data science, or related field; advanced degree or JD a plus but not required in lieu of practical engineering experience.

We offer a competitive compensation and benefits package. Please submit your resume, cover letter and salary requirements.  EOE

#LI-LS1