Hi, my name is
Derrick Lin
Developer · UCLA Med Student · Building tools at the intersection of technology and healthcare
About Me
I'm a medical student at the David Geffen School of Medicine at UCLA with a background in computer science and neurobiology from UC Irvine, where I graduated Phi Beta Kappa and Magna Cum Laude.
Previously, I spent 4+ years in UC Irvine's BCI Lab where I designed custom PCBs, built real-time neural signal visualization software, and engineered hardware for a brain-controlled walking exoskeleton. That work earned First Place at the IEEE EMBC Young Professional Paper Competition and contributed to 10 publications.
Now I'm focused on health tech — I co-founded Transcend Medical, a startup connecting patients with gender-affirming care providers, and Teddy Bear Hospital USA, a 501(c)(3) non-profit that brings interactive medical play to elementary schools, reducing clinical anxiety in children while promoting health literacy. I build tools that bridge technology and healthcare — one project at a time.
Education
MD Candidate, UCLA (2028)
Undergrad
BS CS & Neurobiology, UC Irvine
Focus
Health Tech
Startup
Co-Founder & CTO, Transcend Medical
Projects
Doctolingo
Updated Apr 2026Doctolingo is a pre-launch iOS app that teaches US-based med students, residents, and attendings to speak medical Spanish with real patients. Record yourself saying a clinical phrase — record onset, describe the pain, ask about allergies — and get a word-level heatmap plus per-phoneme accuracy scores powered by Apple's on-device SFSpeechRecognizer (free, offline transcript), a custom Castilian G2P for on-device phoneme comparison, and Azure Pronunciation Assessment for cloud-grade per-phoneme scoring. Pairs that with simplified Leitner SRS (3 boxes), side-by-side native-speaker vs your-attempt playback, per-phrase trend sparklines, a streak counter, and daily goal — all bundled offline-first after a single download.
Key Features
- On-device transcript via SFSpeechRecognizer (free, offline, no marginal cost)
- Per-phoneme accuracy scores via Azure Pronunciation Assessment REST
- Custom Castilian Spanish G2P (ch→tʃ, ll→ʝ, rr→r, ñ→ɲ, ce/ci→θ, ge/gi→x, …) for on-device phoneme scoring
- Word-level green/yellow/red heatmap per attempt
- Simplified Leitner SRS (3 boxes; ≥80 promotes, <80 demotes to Box 1)
- Side-by-side playback: native speaker vs your own recording, independent state
- Per-phrase trend sparkline (last 20 attempts, live-refreshed on each score)
- Streak counter + configurable daily goal with haptic feedback on score arrival
- Fully offline after first launch — ~200 pre-generated WAVs bundled in app
- Pre-launch, private TestFlight beta
ChartLens
Updated Feb 2026Paste clinical notes — H&Ps, progress notes, discharge summaries, consult notes — and get a structured patient overview powered by LLM extraction. ChartLens uses Google's langextract library with few-shot prompting to identify diagnoses, medications, labs, vitals, imaging, procedures, allergies, and social history, all grounded to source text with character-level intervals. Supports Gemini, OpenAI, and local Ollama models. Features automatic PHI de-identification using Microsoft Presidio with custom MRN and age-over-89 recognizers before sending notes to cloud providers, with full re-identification of extracted data afterward.
Key Features
- 10 extraction classes: diagnoses, medications, labs, vitals, imaging, procedures, assessments, allergies, social history, timeline events
- Source grounding — every extraction linked to original text with character intervals
- PHI de-identification via Presidio with custom MRN and age >89 recognizers
- Multi-provider support: Gemini, OpenAI, and local Ollama (no data leaves your machine)
- Tabbed views: Summary, Problems, Medications, Labs, Timeline, Source, and PHI highlights
- Automatic re-identification and offset remapping after extraction
- Docker Compose deployment with optional local Ollama service
Stroke Localizer
Updated Feb 2026An interactive web app for medical trainees to practice neurological stroke localization. Select from 60 symptoms with per-symptom laterality, use the body diagram for quick motor/sensory selection, score the NIHSS, or type a free-text clinical description and let AI parse it into structured findings. A rule-based engine with 57 mapping rules identifies the affected vascular territory, visualized on a rotatable 3D brain, multi-plane 2D SVGs, and a Circle of Willis artery diagram — all with bidirectional hover interactions. Includes named syndrome detection, H&P note generation, a territory explorer, and an integrated Obsidian clinical knowledge base.
Key Features
- 60 symptoms across 7 categories with per-symptom L/R laterality and body diagram quick-select
- 57 mapping rules across 13 vascular territories with confidence scoring and named syndrome detection
- AI-powered free text parsing — type natural language, Claude or OpenAI extracts structured symptoms
- Interactive 3D brain (React Three Fiber) and multi-plane 2D SVGs (axial, coronal, sagittal)
- Circle of Willis artery diagram with hemisphere-aware occlusion markers
- NIHSS score calculator with automatic symptom mapping
- H&P note generator, territory explorer, and clinical reasoning panel
- 63 unit tests covering localization, syndrome detection, NIHSS, and reverse mapping
Transcend Medical
Updated Feb 2026As Co-Founder & CTO, I'm leading the development of the first centralized network of gender-affirming providers — offering everything from hormone therapy and facial gender-affirming surgery to psychiatry, urology, fertility care, and more. The app features real-time messaging, smart scheduling, verified provider search, and The Oracle AI assistant trained on peer-reviewed literature. Available on iOS and Android.
Key Features
- Real-time messaging between patients and providers
- Smart scheduling with color-coded calendar and integrated provider profiles
- Provider search with filtering and reviews across every specialty
- The Oracle AI — 24/7 assistant trained on peer-reviewed literature
- Available on App Store and Google Play
- Cross-functional collaboration with clinical advisors
Photographic Consistency Analyzer
Updated Feb 2026A comprehensive computer vision pipeline that detects and quantifies inconsistencies in lighting, color, composition, and quality between medical image pairs. Uses MediaPipe face mesh with 468 3D landmarks for detection and alignment, 6DRepNet (PyTorch) for 3D head pose estimation, and Signal Detection Theory-based scoring with calibrated JND thresholds (NIST FRVT-conformant 5° pose, CIEDE2000 color). Includes an automatic neck-only fallback achieving 93.9% overall success rate (306/326 pairs from 67 studies). Features a full-stack web app, CLI, and Python API.
Key Features
- Multi-domain analysis: lighting, color, composition, pose, quality
- MediaPipe face mesh (468 3D landmarks) + 6DRepNet head pose estimation
- Automatic neck-only fallback with 75% recovery rate when face detection fails
- 93.9% success rate — 306/326 image pairs analyzed across 67 studies
- SDT-based scoring with calibrated JND thresholds (NIST FRVT, CIEDE2000)
- Full-stack web app with real-time batch processing via WebSocket
- CLI, Python API, and export to CSV, JSON, and PDF formats
Experience
Co-Founder & CTO
Transcend Medical · Los Angeles, CA
Mar 2025 – Present
Leading development of a mobile platform connecting patients with vetted, gender-affirming care providers. Designed core features including real-time messaging, scheduling, and provider search. Launched beta at UCLA.
Research Specialist
Brain Computer Interface Lab, UC Irvine · Irvine, CA
Jul 2023 – Jul 2024
Designed custom PCBs for a 60-channel neural recording headstage (90% cost reduction). Built .NET UI for real-time 15kHz neural signal visualization. Engineered BCI hardware enabling neural control of a walking exoskeleton. Collaborated on FDA IDE thermal testing.
Co-Founder
Teddy Bear Hospital USA · Irvine, CA
Nov 2022 – Present
Co-founded America's first Teddy Bear Hospital chapter as a 501(c)(3) non-profit. Elementary school children bring stuffed animals to mock hospital stations, building positive associations with healthcare. Partnered with UCI, CHOC Pediatric residents, and expanded to UCLA, ASU, Georgia Tech, and UC Berkeley.
Undergraduate Researcher
Brain Computer Interface Lab, UC Irvine · Irvine, CA
Nov 2020 – Jul 2023
Differentiated iPSCs into neurons and developed automated MATLAB image processing pipeline for ~100,000 cells. Optimized neuronal co-cultures on microelectrode arrays for electrophysiology studies.
Publications
10 peer-reviewed publications · View on Google Scholar →
Perception of Brain-Computer Interface Implantation Surgery for Motor, Sensory, and Autonomic Restoration in Spinal Cord Injury and Stroke
D Lin, T Tran, S Thaploo, JGE Matias, K Pixley, Z Nenadic, AH Do
arXiv preprint, 2507.11572 · 2025First Author
Real-Time Brain-Computer Interface Control of Walking Exoskeleton with Bilateral Sensory Feedback
J Lim, PT Wang, WJ Sohn, D Lin, S Thaploo, L Bashford, D Bjanes, A Nguyen, H Gong, M Armacost, SJ Shaw, S Kellis, B Lee, D Lee, P Heydari, RA Andersen, Z Nenadic, CY Liu, AH Do
arXiv preprint, 2505.00219 · 2025
Early feasibility of an embedded bi-directional brain-computer interface for ambulation
J Lim, PT Wang, WJ Sohn, C Serrano-Amenos, M Ibrahim, D Lin, S Thaploo, SJ Shaw, M Armacost, H Gong, B Lee, D Lee, RA Andersen, P Heydari, CY Liu, Z Nenadic, AH Do
IEEE EMBC 2024, pp. 1-5 · 2024
Pruning functional connections in human induced pluripotent stem cell derived neural networks
S Thaploo, D Lin, GJ Brewer, AH Do, Z Nenadic
IEEE EMBC 2024, pp. 1-5 · 2024
A simplified method for long-term maintenance of human induced pluripotent stem-cell derived neural networks
S Thaploo, D Lin, Y Li, M Vu, G Brewer, Z Nenadic, AH Do
IEEE EMBC 2024, pp. 1-4 · 2024
Development of highly sensitive, flexible dual L-glutamate and GABA microsensors for in vivo brain sensing
SS Chu, HA Nguyen, D Lin, M Bhatti, CE Jones-Tinsley, AH Do, RD Frostig, Z Nenadic, X Xu, MM Lim, H Cao
Biosensors and Bioelectronics, 222:114941 · 2023
BCI-based Neuroprostheses and physiotherapies for stroke motor rehabilitation
J Lim, D Lin, WJ Sohn, CM McCrimmon, PT Wang, Z Nenadic, AH Do
Neurorehabilitation Technology, 509-524 · 2022
Defining Outcomes in Facial Gender-Affirming Surgery: A Systematic Review to Inform Core Outcome Set Development
MN Miller, D Lin, S Rabinovich, G Airth, S Rainsbury-Silva, R Canfield, S Fadich, K Shariati, JP Bradley, JC Lee
Preprints · 2025
The Anatomical Breast Burden (ABB) Model: A Schnur Scale Alternative for Identifying Need for Therapeutic Reduction Mammaplasty
E Jolkovsky, MN Miller, W McClain, S Rabinovich, AD Barkhordarzadeh, D Lin, S Piva, A Lerner, GC Slack
Aesthetic Surgery Journal Open Forum, ojaf168 · 2025
Long-Term Maintenance of Human Stem Cell-derived Neural Networks for Electrophysiology Studies
S Thaploo, D Lin, P Wang, A Appajodu, P Thurgam, A Baig, L Vargas, Y Li, G Brewer, Z Nenadic, AH Do
IEEE EMBC 2025 · 2025
Skills & Tech
Languages
Frameworks & Libraries
Tools & Platforms
Hardware & Domain
Get in Touch
I'm always open to discussing health tech projects, new ideas, or opportunities to collaborate. Feel free to reach out!