US-Based Contract Managers, Corporate Paralegals & eDiscovery Specialists — Designing Evaluations

We build evaluations that measure whether AI systems can perform legal operations work to the standard a real matter demands. We are engaging experienced US-based practitioners to design and review them. Why we are specific about US practice The work turns on how this is actually done in the US: contracting conventions and playbook positions used by US legal teams, state and Delaware corporate filing practice, and eDiscovery under the Federal Rules. Process from another jurisdiction does not transfer cleanly, and a task built on the wrong assumptions produces a confidently wrong answer that nobody catches. We are therefore engaging practitioners based in the US with US in-house, law firm or ALSP experience. Admission is not required — this posting is for the operational side of legal, not for counsel. We have a separate posting for US-qualified lawyers. What the engagement involves You would take a category of work you have done — an obligation extraction, a clause deviation review across a portfolio, a renewal and auto-renew calendar, an entity register reconciliation, a closing checklist, a custodian list, a privilege log — and build an evaluation task from it: - Construct a representative fact pattern and the documents a practitioner would receive - Specify what a correct answer must establish, and what a plausible but wrong answer looks like - Review tasks written by other practitioners and identify anything that would not survive a real matter under deadline The second and third points are the substance of the work. Defining a defensible wrong answer is considerably harder than defining a right one, and it is the part most evaluations get wrong. Who this suits - Contract managers and CLM specialists — obligation extraction, renewal and auto-renew calendars, negotiating against a playbook, clause deviation reporting across a portfolio, third-party paper risk scoring. Ironclad, Icertis, Agiloft, DocuSign CLM or similar - Corporate paralegals and company secretarial — entity and subsidiary registers, filing deadline dockets, signature verification, closing checklists, board resolutions and minutes - Litigation support and eDiscovery specialists — hold notices and custodian lists, review coding and privilege logs, deposition prep chronologies. Relativity, Everlaw, Nuix or similar We are looking for people with real matter experience rather than coursework, and the ability to write a task precisely enough that another practitioner could follow it exactly. Terms - Paid per completed task, $50–$250 depending on complexity and length - Ongoing rather than one-off, with steady volume for practitioners whose work holds up in review - Entirely asynchronous and in writing — no meetings, no calls, no fixed hours - Around 10 hours a week if you want consistent volume; there is no minimum Confidentiality You would author from representative examples you construct yourself — never real contracts, client matters, custodian data, or privileged material. We will not ask you to reproduce anything you have worked on, and any task that appears to draw on a real matter is rejected. Please answer the questions below. We read every response. The question about the mistake a junior practitioner makes on this work tells us considerably more than a list of years and titles.