Senior Corporate Law Expert

You'll be the go-to legal expert for a research team at a frontier AI lab, working directly with their legal research lead. Your job is to define what a model needs to get right in corporate and transactional law, then hold it to that standard. You'll build and review legal eval tasks, critique the model's answers to real legal questions, set the quality bar for the team's legal data, and help vet the other lawyers who contribute.

The material comes from the work you've done in practice: structuring an acquisition, reading a license for what it grants, running down a diligence list, advising a board on a governance question, spotting the securities issue in a financing. You already know where current models give a confident, polished answer for the wrong jurisdiction or the wrong deal posture. Here, that instinct becomes the test. The research team will lean on your judgment to decide which legal problems are hard for models and how to measure them.

This is a remote, 4-month engagement to start, open to lawyers anywhere in the US. Seattle, San Francisco or New York is a plus. It's designed to run alongside a full-time job. It comes with real ownership over how this lab handles legal work.

What you'll do

• Build legal eval tasks from realistic transactional scenarios, such as asset vs. stock vs. merger structuring, indemnification caps and escrows, and change-of-control and anti-assignment provisions, each with an answer key a senior deal lawyer would sign off on
• Critique the model's answers to corporate law questions and pinpoint where the reasoning breaks, such as a misread license grant, a missed third-party consent in diligence, or Delaware fiduciary-duty doctrine applied to the wrong fact pattern
• Pressure-test the model on securities and capital markets questions, including private placement exemptions, disclosure obligations, resale restrictions, and board approvals for a financing, and judge whether its answer would hold up with a real client
• Write and document the quality bar for legal data: what a correct, complete answer looks like, which omissions are disqualifying, and how to grade answers that are close but wrong
• Review the tasks and critiques of other legal experts, help vet new contributors, and flag where their work falls short of the standard
• Help the research team map the domain, including which corporate, commercial, and cross-border problems are hard for models, why they're hard, and how to test them
• Work through the data line by line, at the level of detail you'd bring to a closing checklist

Who we're looking for

• You have a JD, you're admitted to practice in the US, and you have 5+ years of practice with meaningful transactional or corporate work
• You've worked as in-house counsel, ideally after starting at a large law firm, and you've seen a range of companies, from larger organizations to startups. A top law school or firm is a plus but not required
• You have broad corporate fluency across M&A, governance, commercial contracts, licensing, and financings, plus enough command of adjacent areas like employment, tax, IP, and privacy to know when an issue belongs there. Your practice spans the field rather than one niche
• You already use AI tools in your own legal work and can name specific places where they get the law wrong
• You want to do the work yourself: reviewing data, writing tasks, and being thorough about the details. Managing other contractors or doing business development isn't what you're after
• You're curious about how models are trained and evaluated, and you want to understand the mechanics
• You're authorized to work in the United States and can commit to 10-35 hrs/week for 4 months

Details

• Pay: $300 / hr
• Workload: 10-35 hrs/week
• Duration: 4 months
• Start: Immediately
• Type: W-2 Contract

Massive (Do a Huddle, Inc.) runs a short intro call and interview and names the client before you commit. Never pay to get work.