Fractional Legal AI Engineer

100% Remote Full-time Open now

Role Who you are

Hands-on experience building and shipping AI-assisted workflows — not evaluating tools or writing specs, but actually standing systems up. Be ready to walk through something you built from prompt to production
Direct, working proficiency with Claude or comparable LLMs: prompt engineering, agentic workflows, and iterating on outputs until they're reliable. This is the core technical skill for everything we're building
Integration experience — you've wired tools together using APIs, MCPs, or similar patterns, and you know what breaks and why. Coordinating with engineering or IT teams to scope and execute those connections is second nature to you
Comfort operating in ambiguity. You can take a business problem and context from attorneys or legal ops — not a technical spec — and turn it into a working system. You know how to ask the right questions, fill in gaps with judgment, and move forward
Comfort working within internal technology governance processes. You know how to communicate with an IT or AI team on tool approvals, data access, and security reviews — and you don't try to go around them. You build within the guardrails, and you help the team understand why those guardrails matter
Bonus: Experience working alongside attorneys: you understand contracting workflows, matter intake, research cycles, and how legal teams make decisions

What the job involves

We're looking for a Legal AI Engineer for a fixed engagement to build foundational legal team AI infrastructure
You'll design and stand up these systems, document them thoroughly, and hand them off to the team in a state where they can be maintained and iterated on internally
You'll provide services to a legal team that is already building its own AI tooling and infrastructure through Claude
Your engagement will be with the Senior Manager of Legal Contracting Operations & AI Systems
Your project will be to add the technical depth that turns promising experiments into robust, shared systems, and to deliver a coherent stack of tools and workflows
During this engagement, you'll design workflows that are fast, measurable, and safe based on specifications from legal, legal ops, and cross-functional stakeholders
You'll leave behind documented systems and run books so the team can maintain and evolve what you've built
This engagement is equal parts legal infrastructure + applied AI, ideal for a hybrid legal/technical practitioner who can translate legal work into scalable systems and then make those systems real

Connect the AI Stack:

Take the skills, workflows, and dashboards that individual team members have already built in Claude and package them into shared, maintainable assets the whole team can use
Wire together current AI tooling — build MCP integrations so data flows where it's needed without manual handoffs; engage with Gusto's AIT team on connections where native integrations don't exist
Further operationalize AI tools and integrations as agreed upon

Build a Legal Repository:

Stand up a structured, queryable knowledge layer fed by the team's existing research, positions, precedents, and playbooks — one that can be searched by any team member, can be invoked by automated workflows, and continuously ingests new information as the team works in the tool

Implement AI Workflows:

Working from priorities set by legal ops and attorneys, implement AI-assisted workflows for the team's highest-volume use cases — likely including contract first-pass review, matter intake and routing, litigation operations to enable intake, routing and response drafting for subpoenas, demand letters and notices and prompt libraries for recurring legal tasks

Document and Hand Off:

Deliver a maintenance runbook for every system built and train designated team members on administration, so nothing requires the contractor to maintain it after the engagement ends. Establish adoption baselines so the team can measure progress and know what to iterate on next

Engagement Details:

Apply tot his job Apply To this Job

You might also like