Journey to the Frontier.
In four weeks you work with an experienced legal quant to build a tool for a real problem in your practice. You do not need a finished idea or a technical background.
45 lawyers from law firms and in-house legal teams have completed the Residency across three cohorts. Some firms sponsor places. Many lawyers pay for themselves.
“The LegalQuants Residency is part of that model: help some of our most AI-forward team members better learn the ‘art of the possible’, which ultimately makes LASER’s job easier and gets us to client impact faster.”
Andrew SimonChief AI & Innovation Officer, WeilAlso published by Weil →“I think the program itself really opened my eyes to a different way of thinking about solving legal problems, helping clients in creative ways and understanding really what’s possible now in this new paradigm.”
Jason HogganSpecial Counsel, Sheppard Mullin · Cohort 3When someone asks the most impressive thing you’ve done with AI, you’ll have a concrete answer.
You’ll be matched on practice area, seniority, timezone, and the problem you want to solve. Some residents arrive having never opened a terminal. Others are six months into a codebase they can’t verify. Either way, what you need is a senior lawyer who builds, telling you what’s right and what isn’t.
An optional week to get ready, then four weeks to build your project.
You’ll tell us where you are, what tools you currently use, and where work is slowing you down. We use that to assess the fit, identify a promising project, and brief your mentor before the residency begins.
Five days on how Claude Code and Codex actually work, in your own hands. One module a day: a short recorded walkthrough, a live clinic, and an exercise a mentor checks. Not taking it? You use the week for setup, with a step-by-step guide, a setup lab and tech support.
You scope your project with your mentor on day one, then build it. You have nine scheduled one-to-one sessions with your mentor in the first three build weeks, and you can message your mentor between sessions. You can also bring problems to group AI clinics, attend office hours, and compare approaches with the rest of the cohort. The last week is for finishing, submitting and mentor review.
You’ll show your project to the cohort and answer questions: how it works, what you learned, what you’d improve next. To graduate, you then sit LQ Assess. Graduates earn the AI Residency Graduate credential.
Time commitment: 10–15 hours per week · Typical build time: 1–2 hours per day · Foundations: 2–3 hours per day
See the cohort schedule: November – December 2026
Week one is the optional Foundations week: a live clinic each weekday at 6am PT · 9am ET · 2pm London · 3pm CET · 10pm SG/HK, recorded. Then six webinars on Tuesdays and Thursdays at the same time, all recorded. Each is led by a leading practitioner within the LegalQuants community. Individual topics are confirmed closer to each date and published on your schedule once set. Your mentor 1:1s are 60 minutes on Mondays, Wednesdays and Fridays in the first three build weeks, at times you and your mentor set. Office hours run in small groups, at times set from the cohort’s timezones once the room is known.
A practical week to get comfortable with the tools before you start your project.
Our last cohort told us the same thing. The build was the best part. The hardest part was everything before it: getting set up, knowing what the agent had actually done, and seeing how the pieces of a real tool fit together. Foundations is that week, taught properly, before the build starts.
What each day covers, who it’s for, time and price
It is not a lecture series. Every day you do one thing with your own hands, post the result, and a Foundations mentor checks it. On Mon 9 November you start your own build on day one.
What actually happens when you type: you, the agent, the model, its tools, your folder. Plan mode: read the plan, change it, then say go.
The file your agent reads every session, and what belongs in it. What a skill is: install one, then write your own in ten lines.
Why long sessions go wrong and how to notice. Starting fresh from a handover note. Which model for which job, what it costs, and what to do at your limit.
What a tool call is. Build a ten-line MCP server, connect it, and watch the agent use it.
Checking an agent’s work without reading code. Undoing a change. A fresh session, then the other agent, reviewing the work. Claude Code and Codex side by side.
Anyone who has mostly used AI in a chat window. Anyone who has built something but couldn’t say exactly what the agent did. Anyone who wants the build weeks spent on their build, not on their setup.
If you already work in Claude Code or Codex daily, keep a repo on GitHub, and have deployed something you built, the four build weeks alone are designed for you.
About two to three hours a day for five weekdays: a 20 to 30 minute walkthrough, a one-hour live clinic, and the exercise. Clinics are recorded. $980, on top of tuition.
What past cohorts built.
Where each resident started, what they shipped, and who mentored them.
A practice-innovation lead from a law-librarian background — comfortable with code, and wanting the practice experience his lawyer peers already had.
A computer-vision model, built and trained from the ground up, that pairs the images in past IP cases with the court’s analysis. On his benchmark it scored 84.23%, against 74.58% for Fable and 59.36% for Astra; first place in the cohort.
“What I think the cohort did really well is that it allows people to find something that they’re innately curious about… and to grow through that.”
Justin🇺🇸 Practice innovation lead
Emily🇺🇸 MentorWant to hear it straight from someone who did it? Message Jamie on LinkedIn for an intro to a past resident.
What our residents say.
Every graduate of our last cohort presented a tool they built from a problem in their own practice. Here is what they said at the closing feedback session.
“The secret sauce here is that it is a one-on-one relationship, and that you are building and you’re learning by doing.”
Monica · Cohort 3Special Counsel · AmLaw 50 firm
“From a time-proportion perspective, more value added to my life than my law school degree.”
Sandeep · Cohort 3Senior Counsel · Global payments company
“I started from scratch. So everything I’ve learned, I’ve learned on this course, and you can see where I’ve gone from zero to having a working product.”
Kathryn · Cohort 3VC & Digital Assets Counsel
Posted publicly by graduates, and by a graduate’s firm, in the week after Demo Day.
What’s included.
Lawyers from AmLaw100, Magic Circle and Fortune 500 in-house teams, all mid-build at the same time as you. Different projects, same problems.
Nine scheduled 1:1s with a senior lawyer in your practice area, plus your mentor on Slack between calls. Dan (tax) mentored Mark (tax). Parth (patents) mentored John (IP litigation).
Bring the thing that broke on Tuesday. Plus every past cohort’s session, so the answer usually already exists.
Access to specialists for focused questions outside your mentor’s area.
The community chat, the Residency alumni group, and local meetups.
Our assessment of how you work with AI. The residency ends with Demo Day and LQ Assess. Your build and your result go on your legalquants.com profile, to share if you choose.
You build in your own Claude, Codex or other AI subscription, so what you make keeps working after the month. You also get LQ Brain, answers drawn from LegalQuants members’ own work.
It plugs into the AI you already use. Install it once and your Claude Code or Codex knows what is due this week, what your mentor sent, and how to hand work in. On the web, you and your mentor share one room for the plan, the calls and the build log. Every session is recorded, with the Mirror, the reading, LQ Brain and the Foundation Tapes podcast alongside. You keep access after the residency.
Every mentor practised law. Every one of them was vetted by LegalQuants.
Every mentor has deep legal domain expertise, frontier AI knowledge, and has built software that scaled. They understand what good looks like for legal work, so you build to a lawyer’s standard, not a demo standard. You’ll be matched with someone whose experience is relevant to your work.
LegalQuants has enough senior lawyer-builders to match you on practice area, time zone and seniority at the same time.
Practice areas we cover: privacy · funds · finance · tax · IP/tech · litigation · arbitration · restructuring & insolvency · trusts and estates · capital markets · in-house legal operations · and many more.
Nine time zones, UTC−7 to UTC+12
Co-Founder, LegalQuants · ex-Clifford Chance funds lawyer
Solo-built internal tools that scaled to 18 Clifford Chance offices in 10 months · kickstarted the vibe-coding movement among lawyers
Rajal · FundsFiona · FundsJerome · FinanceFelix · Legal engineering
Co-Founder, LegalQuants · Senior Associate, Technology · global top-20 firm
Built the Global AI Regulation Tracker solo · Lawyers Weekly 30-under-30 Tech Lawyer of the Year, 2023
Peter · PrivacyAverell · CryptoJoe · Tech lawKathryn · Venture capital
Senior Counsel, Group HQ · Masdar
ex-Shearman & Sterling, Arbitration (Paris) · ex-GC, EGA Africa
Gillie · PrivacyBill · Tech GCJose · CorporateJoyce · Employment
Senior Legal Counsel · Toyota Tsusho Asia Pacific
71+ open-source repositories
Kenny · TechAshley · Debt financeRoger · Privacy
Joey · Real estateAgatha · In-houseErica · Governance
Private Client, Trusts & Estates · AmLaw 50 firm
Trusts and estates for private clients
Tamur · Managing partnerElias · Private wealthSpencer · Legal AI
Sean · LitigationHank · CLOMonica · Legal training
Solicitor Advocate & commercial litigator · UK litigation firm
Builds local, self-hosted AI for legal work
Judith · EntrepreneurAlex · ArbitrationJason · White collar
Manager, Credit Management & Debt Recovery · Fletcher Building
Former insolvency lawyer
Alexis · Corporate counselPetra · Political law
Associate, TMT · leading Central European firm
Dual LL.M., King’s College London + Bocconi · École 42
Jayne · CorporateSandeep · Fintech
Senior Associate, Restructuring & Insolvency · Magic Circle firm, seconded to its Global Innovation Office
Inside a Magic Circle innovation function
Fatema · GovernmentChris · Commercial
Special Counsel, Litigation & Arbitration · AmLaw 50 firm
Yale Law · Second Circuit clerkship · 15+ years litigating securities-fraud cases
Bradley · Finance
Founder · Searchlight
ex-Orrick patent litigator · Harvard Law · Berkeley EE/CS · Federal Circuit + Delaware clerk
John · IP litigation
Mark · Tax
Tax lawyer & Chartered Accountant · IFM Investors
Builds cross-disciplinary tools for tax, legal and finance workflows · co-author of the AgentArcEval paper · ex-Luna (AI tax copilot)
Dipti · Legal tech
Corporate lawyer · AI builder · ex-Sidley Austin
Five years advising listed companies on M&A and capital markets, then pivoted to product · HKEX size-test agent, Listing Rules RAG, HK judgment summariser
Rita · Practice systems
Senior Counsel · Clipboard Health
Eight years a federal prosecutor, then in-house at Facebook · built the open-source Legal Risk Pricer, a multi-agent Claude skill that turns a contract position or regulatory change into a quantified risk assessment
Abdul · M&AAssociate General Counsel · AI-native venture studio
Built Privilege Sentinel, an open-source privilege-risk assessment tool for legal AI prompts · now designing multi-agent systems to rebuild in-house legal operations from the ground up
Justin · InnovationThis is the current bench, not a fixed roster — mentors join and occasionally step back between cohorts. Being pictured here isn’t a placement guarantee; you’re matched to whoever fits your practice area, timezone and seniority best.
The LegalQuants community.
Lawyers from AmLaw100, Magic Circle, and Fortune 500 in-house teams. One chat community, 1,000 messages a week, some of the most AI-native lawyers anywhere. Once you graduate, you’re in. After that, your stuck point is usually someone else’s solved problem.
More from the community
Included: 1:1 mentoring, group sessions, learning platform, invitation to LQ Assess. Bring your own AI subscription.
We review your enrolment within 48 hours of payment. If we don’t think the fit is right, you get everything back, no reason required.
Employer reimbursement supported: invoice, course summary, and certificate of completion. Graduates also receive a verifiable AI Residency Graduate credential.
If you’d rather talk it through first, book a call with Jamie — or ask for an intro to a past resident or a mentor in your practice area.
Asked before joining.
What do I actually learn?
You become an AI-native lawyer by working next to ones who already are. The hard skills come with it: meta-prompting, workflow mapping, document ingestion pipelines, delivering work as HTML, building your own tools, and verifying AI’s work instead of trusting it.
What will I actually make?
One of four shapes, sized to your practice. A workflow: a repeatable Claude Code process for a matter you handle weekly. A tool: an internal app your team can actually use. A capability: running an agent autonomously for hours, or standing up an MCP endpoint. A system: the practice itself, redesigned. The project is the tangible result of a deeper change in how you work.
Do I need a technical background?
No. Most residents had never opened a terminal. There are prerequisites before the build — installing the tools — and a step-by-step guide plus tech support for exactly that. If you’d rather arrive already practised, that is what Foundations is for.
What’s the time commitment?
10–15 hours a week in the build weeks: your 1:1s, about three hours of office hours and webinars, and 1–2 hours a day on your build. Foundations, if you take it, is about two to three hours a day for its five days.
Can I use firm documents? What about confidentiality?
We can sign an NDA, and we keep a library of open and synthetic documents so you never have to touch client data. Nothing you build is public unless you decide it is.
Should I take Foundations?
Take it if most of your AI use is in a chat window, or you have built something but couldn’t say exactly what the agent did. Skip it if you work in Claude Code or Codex daily and have deployed something of your own. Your Mirror result will give a recommendation. The four build weeks alone still take you from prompting to building; Foundations means those weeks go on your build rather than on the basics.
Do I pick my mentor?
We match you — practice area, seniority, timezone. Dan (tax) mentored Mark (tax). Jamie (funds) mentored Rajal and Fiona (both funds). Parth (patents) mentored John (IP litigation).
What’s the refund policy?
Two paths. We review your enrolment within 48 hours of payment and, if we conclude the fit isn’t right, you get a full refund, no reason required. Or you cancel — one email to admin@legalquants.com — 50% refund up to 25 October 2026; after that, no refund except force majeure.
More questions (12)
I already use ChatGPT. What do you mean by “build”?
Most lawyers bolt AI onto the way they already work: receive a task, open ChatGPT, paste in, read out, copy back — and start from scratch on the next matter. Building means rebuilding the work around AI: your own tools in Claude Code, agents that run while you’re doing something else, processes you keep and improve.
Am I building alone?
No. The cohort includes lawyers from major firms, in-house teams, and legal technology companies — everyone on a different project, but the problems overlap: repetitive review, scattered information, slow handoffs. The group is where you see how others approach similar problems, get unstuck between mentoring sessions, and test ideas with people who understand legal work. Once you graduate, you join the LegalQuants community, with the alumni group and local meetups.
I don’t have a build idea yet. Is that a problem?
No. Picking what to build is half the work, and the assessment and setup weeks are partly for finding it with your mentor. A good test: the thing you’d use yourself, daily. If you wouldn’t, neither would anyone else. You don’t need an app, either; a workflow, a document pipeline, a monitoring routine, or a repeatable capability all count.
I’ve already enrolled. Can I add Foundations?
How long is it?
Four weeks of building, plus an optional Foundations week before it. Week 1 (Nov 2 – 6) is Foundations for those who take it, and setup for everyone — git, GitHub, Claude Code or Codex, Vercel, and the reading. Weeks 2–5 (Nov 9 – Dec 4) are the build. Demo Day is December 6.
What if work slams me mid-cohort?
You can defer to the next cohort and make up the remaining time. Once, not indefinitely.
My firm blocks Claude Code. Is this still worth it?
Yes, and this situation is common in the cohort. You learn the frontier on your own machine and your own time, and that knowledge transfers: residents in locked-down firms end up knowing what the tools can actually do. We calibrate to what you’re allowed to run.
Couldn’t I learn this from a bootcamp or YouTube?
The difference is the person across the table: a senior lawyer in your practice area who knows what good legal work looks like. Generic courses can’t see your bottlenecks. Your mentor has usually had them.
I’m a lawyer, not a builder. Is this for me?
The aim isn’t to make you a developer. It’s leverage: using AI to build the tools that do your work.
What happens after I graduate?
You graduate once you have presented at Demo Day and sat LQ Assess. You then join the LegalQuants community and the Residency alumni group, and your build and your result go on your legalquants.com profile, to share if you choose.
What support is there after the residency?
The LegalQuants community chat, the Residency alumni group and local meetups.
Why LegalQuants and not someone else?
The bench: enough senior lawyer-builders to match you on practice area, timezone and seniority at once. And the mentors all know each other — you’re joining their community for the month.
You came here asking whether you’re doing it right.
Four weeks from now, you will leave with a clearer sense of what AI can do in your practice, and feedback from a mentor who understands the work. If you would like to build that with us, join Cohort 4.
Reserve your seat in the first cohort of 2027.
$500 holds your place and today’s price. It’s credited in full toward tuition, and transfers across 2027 cohorts if your plans move. We’ll write the moment the cohort’s dates are set.
Not ready to commit? Leave your details and we’ll write when the next cohort’s dates are set.