A program director sees a moment worth teaching. They speak for under a minute — and Sherpa turns it into signed, competency-mapped documentation. The growth memory forgets is finally captured.
Real development happens daily — in a hallway, after a code, mid-rounds. But in today's systems, recording a win is as onerous as recording a concern. So concerns get documented, wins evaporate, and oversight only ever sees the problems.
Defensibility means the negative is always preserved — and, being negative, it lingers on the record long after the moment has passed.
Too much friction, and nothing that compels it. The wins fade — by the next committee meeting, the good is simply gone.
Fix the capture once — and both the trainee and the institution win.
The review becomes a conversation again — not a form. The program director stays in the moment; the write-up takes care of itself. Praise or concern, the same quick flow.
Tap and talk for under a minute on any device, before the moment evaporates. No note-taking, no form.
It resolves the subject and the ACGME competency, drafts a third-person report, and reads it against national data.
Evidence now, the committee queue, or both. You review and sign every word before anything is recorded.
Works mid-conversation, on any device, across six ACGME-mapped review types.
Third-person prose by ACGME domain, ready in seconds — not a 30-minute write-up.
One-click I-SMART goals and structured remediation, pulled straight from the transcript.
Ambient committee capture, milestone ratings, and a print-ready report for every cycle.
Sherpa is not a second system to maintain, and it does not replace New Innovations or MedHub. A moment is captured and ratified in Sherpa; the finalized result flows into your system of record.
Present-first, on any device. Spoken in under a minute — no note-taking.
Human-approved, evidence-backed, competency-mapped. You sign every word.
Flows into New Innovations or MedHub. Coordinators stop re-typing.
Speech becomes structured observation — subject, competency, and suggested level.
Each moment is scored against national milestone data to separate curricular signal from individual.
The committee ratifies; the signed result reconciles into your existing system of record.
The same sixty-second moment answers a different need at every level of the program — from the resident on the floor to the institution's oversight.
Growth is recognized the day it happens, not months later — and repair begins the moment a gap appears.
Stay in the conversation. The 20–30 minute write-up becomes a signature, and downstream documents grow from the same capture.
Mid-cycle development becomes visible across every program — a live view instead of a twice-a-year milestone snapshot.
The hallway review is built for the phone already in your hand. Native apps for iOS and Android are on the way — capture a moment between cases, anywhere it happens.
Sherpa handles assessments of named physicians in training. The controls below are engineering facts about how the system is built and enforced, not aspirations.
Database region and model-processing locations are fixed per environment and recorded as one immutable residency set. A deployment fails if that trio is missing or inconsistent — the check runs in the pipeline, not in a policy document.
Every cloud service the platform enables must carry a dated, classified entry in a service manifest stamped from the provider's own covered-products list. The estate cannot silently gain a service: continuous integration fails on any gap.
Each institution is a separate identity tenant, and every record read is scoped to the institution that owns it. Sessions are server-side; the browser holds an opaque reference over an encrypted connection, restricted to the paths that need it.
Mobile notifications carry an opaque reference and never clinical content — a phone screen is not a private place. Notifications are also deliberately rare: routine progress does not interrupt anyone's day.
Schema changes execute under a restricted role, so a change that would only succeed with elevated privileges fails in rehearsal rather than in production.
Analysis proposes; a person approves. Every assessment that reaches a trainee’s record has been read and signed by a faculty member, and the official system of record remains the source of truth.
Last updated 2 August 2026. This policy covers gmesherpa.com and the Sherpa GME platform, including our mobile applications.
Math Gambit LLC, 835 N Webster St, Naperville, IL 60563, United States. Questions about this policy: privacy@gmesherpa.com.
Assessment records are held for a sponsoring institution under written agreement. The institution decides what is collected, who may see it and how long it is kept; we process it under their instruction and do not use it for our own purposes.
Names, institutional email addresses and programme roles for people the institution enrols; assessments, learning plans and related notes created by faculty; and technical records — sign-in times, device identifiers used to deliver notifications, and error logs needed to keep the service working.
We do not sell personal information and we serve no advertising. Assessment content is not used to train general-purpose models: our model-processing terms exclude submitted content from training, and no assessment leaves the processing region pinned for your institution.
Requests about an assessment record should go to your institution, which controls it. For anything we hold directly, write to privacy@gmesherpa.com and we will respond within 30 days.
Records are retained for the period set in the institutional agreement and deleted or returned at its end. Hosting and model processing run on Google Cloud in a pinned region. A current sub-processor list is available on request.
Last updated 2 August 2026. Institutional use is governed by the signed agreement between Math Gambit LLC and your organisation; that agreement prevails wherever it differs from this page.
Accounts are issued by an institution to identified faculty, trainees and staff. Credentials are personal and are not to be shared. Sherpa is not offered to the general public and is not for personal medical use.
Sherpa supports educational assessment of physicians in training. It does not diagnose or treat patients, does not direct clinical care, and its output is a draft for a qualified person to review, edit and sign.
Assessments and records created in Sherpa belong to the institution that created them. We claim no ownership and use them only to provide the service under the institutional agreement.
We maintain the service in good faith and give reasonable notice of material change. Availability commitments, support response times and liability are set by the institutional agreement rather than by this page.
Every message reaches a person who works on the product. There is no ticket maze.
support@gmesherpa.com · 312·282·6846. Include your institution and programme, and what you were doing when it went wrong. We reply within 2 business days.
Accounts are created and removed by your institution. For a password reset or a change of role, start with your programme coordinator; we cannot add or remove users on an institution’s behalf.
Report a suspected vulnerability to security@gmesherpa.com. We will acknowledge receipt, keep you updated, and will not pursue anyone who reports in good faith and does not access data that is not theirs.
We're partnering with a small number of programs to shape what Sherpa becomes. If it sounds like it fits how your program already works, we'd love to talk.