Advancing patient care through intelligent medical technology.
Our philosophy
AI Medical Solutions was built on the belief that emerging technologies should serve patients first. Our work is grounded in clinical realities, from telehealth challenges to medical language barriers, ensuring that innovation translates directly into better care. We aim to make healthcare more affordable, accessible, and effective.
Our partnerships
We collaborate with physicians, biotechnology researchers, clinical researchers, and academic medical institutions to ensure every tool meets clinical standards. From NSF grant applications to IRB protocols, our work is guided by medical advisors who help us align innovation with real-world care. These partnerships are the foundation that allows us to refine, validate, and responsibly scale our solutions.
Our tools
Each tool is built to reduce clinician burden while improving diagnostic clarity. All are in active clinical research.
- HealthVision Clinical needMusculoskeletal exams over telehealth, where no one can measure a joint by hand. What it measuresJoint range of motion at the elbow, wrist, hip and knee through the patient's phone.
- Multilingual intake assistant Clinical needPatients with limited English proficiency who cannot describe their symptoms with precision. What it deliversA pre-visit history taken in the patient's language and summarized in English for the physician.
- Facial nerve assessment Clinical needFacial recovery after head and neck surgery, still graded largely by eye. What it measuresRegion-by-region facial nerve function from facial landmark tracking.
HealthVision
Range of motion, measured on a video call.
HealthVision began as a digital goniometer for telehealth patients. It uses the phone's camera and motion sensors to guide the patient through standardized movements, measures joint angles in real time, and builds an objective record a clinician can follow from visit to visit.
It was built after physiatrists described how hard it is to examine a joint accurately over telemedicine. HealthVision quantifies, logs and graphs each patient's movement so that a remote visit can include the objective measurement usually taken by hand in the room.
Elbow, wrist, hip and knee
Sample motionHow a measurement is made
From a phone camera to a number a clinician can chart, in five steps.
-
Capture
The patient's own phone records a guided, standardized movement. No added hardware.
-
Landmark detection
Computer vision locates the joint centers in every frame, with motion sensors for orientation.
-
Error correction
An AI layer checks each measurement before it is accepted.
-
Angle computation
Joint angles are computed from the landmark geometry in real time.
-
Clinician dashboard
Every reading is logged and trended across visits for the care team.
HealthVision specification
- Hardware
- The patient's smartphone: camera and motion sensors
- Protocol
- Guided, standardized movements
- Output
- Joint angles in real time; longitudinal record
- Validation
- Formal comparison against conventional goniometry in preparation
- Research module
- Nutrition tracking
- Intellectual property
- Multiple patent applications filed
Multilingual intake assistant
The patient answers in Spanish. The physician reads the history in English.
Patients with limited English often cannot describe pain with the precision a diagnosis depends on: tingling, burning, radiating. The assistant takes the history before the visit, in the patient's language, follows up where an answer is incomplete, and delivers a structured summary to the physician.
Currently being clinically piloted with Spanish-speaking patients.
- Hola. Soy el asistente de admisión de la clínica. ¿Cuál es el motivo de su visita hoy?
- Me duele la rodilla derecha desde hace tres semanas, sobre todo al subir escaleras.
- ¿Cómo describiría el dolor? Por ejemplo: punzante, ardor, hormigueo.
- Punzante, y por la mañana la rodilla está rígida.
- ¿Toma algún medicamento para el dolor?
- Ibuprofeno, 400 mg, dos veces al día.
- Chief complaint
- Right knee pain, 3 weeks
- Aggravated by
- Climbing stairs
- Quality
- Sharp, stabbing
- Associated
- Morning stiffness
- Medications
- Ibuprofen 400 mg, twice daily
Conversation
A large language model conducts the intake in the patient's language, one question at a time.
Adaptive follow-up
Incomplete answers trigger targeted questions on onset, quality, severity, timing and medications.
Clinical structuring
Answers are organized into a structured history and rendered in English.
Delivery
The summary reaches the clinic's staff dashboard before the patient is seen.
| Patient | Language | Chief complaint | Status |
|---|---|---|---|
| Patient A | Spanish | Right knee pain, 3 weeks | Summary ready |
| Patient B | Spanish | Low back pain radiating to left leg | Summary ready |
| Patient C | Spanish | Shoulder stiffness after fall | In progress |
Facial nerve assessment
Developed in research collaboration with Weill Cornell Medicine, Otolaryngology and Head and Neck Surgery.
Recovery after parotid, skull base and head and neck surgery is still graded largely by eye, and grades vary between examiners. This platform tracks facial landmarks while the patient performs standard movements and reports twelve micro-measurements across six nerve territories: symmetry and amplitude indices for each territory, a composite score, and a localization map for the surgeon. It stores landmark coordinates, never raw video.
Sample symmetry by movement. Select one to watch it measured.
Platform specification
- Landmark model
- MediaPipe Face Mesh, 468 landmarks per frame
- Measurements
- 12 micro-measurements across six nerve territories
- Indices
- Per-territory symmetry and amplitude; composite score
- Dynamic metrics
- Excursion amplitude, closure velocity, regional coordination
- Reference frame
- Midline anchored to stable forehead and nasal landmarks
- Clinician output
- Localization map; CSV and JSON export
- Privacy
- Landmark coordinates stored, never raw video
- Reference standard
- Blinded clinician grading on House-Brackmann and Sunnybrook, in development
Ryan Morgenstern
Our founder
AI Medical Solutions was founded by Ryan Morgenstern with a mission to develop innovative technologies that make healthcare more precise, accessible, and effective.
Ryan graduated magna cum laude from New York University, where he was inducted into Phi Beta Kappa. His research background spans psychophysiological biosensor validation, computational medicine, and digital health innovation.
Ryan is also the author of Medicine and the Artificial Intelligence Age: A Decision Science Framework for Evaluating Digital Health Technologies. Written for clinicians navigating the rapidly evolving landscape of medical technology, the book presents a framework for evaluating when and how emerging innovations can meaningfully improve clinical practice.