Pediatric resident at TU Teaching Hospital, Nepal. Full-time doctor, part-time coder.
I build tools for problems I run into during ward rounds, night duties, NICU shifts, and exam prep. Most are designed for low-resource settings where internet access is unreliable and good clinical references are difficult to access quickly.
A clinical workspace for pediatric residents combining bedside tools, evidence search, and AI-assisted learning.
| Tool | What it does |
|---|---|
| Ask Nelson | AI-assisted revision and pediatric clinical Q&A grounded in Nelson Textbook concepts |
| Ask Harriet Lane | Drug dosing and bedside pediatric reference support using Harriet Lane data |
| Ask Evidence | PubMed-backed literature search with AI-synthesized summaries and real PMIDs |
| DysmorphAI | Face2Gene-style dysmorphology AI for syndrome differentials |
Night Duty
- Resuscitation calculator
- Fluid bolus calculator
- ABG interpreter
- Sodium correction tools
NICU
- Jaundice and phototherapy thresholds
- Neonatal drug dosing
- NICE neonatal sepsis support
- APGAR, Ballard, and Thompson scoring
- UVC sizing
- Weight-for-gestational-age tools
PICU
- Pediatric GCS
- Phoenix Sepsis Score (2024 SCCM criteria)
- Critical care bedside support tools
Ward Round
- Harriet Lane drug dosage browser
- WHO/CDC growth charts
- Pediatric BP percentiles
- Nelson 22nd edition lab reference values
- Developmental milestones
- Procedure guides
PocketMed AI Android
Offline-first clinical assistant. No internet required.
On-device AI for:
- Drug dosing
- Differential diagnosis support
- Lab interpretation
- Acute management guidance
Built around real hospital workflows in Nepal and similar low-resource settings.
Privacy-first architecture. Patient data never leaves the device.
Web-based clinical AI assistant.
Features include:
- Diagnostic support and case reasoning
- Drug dosages, contraindications, and interaction checks
- Clinical knowledge base and disease summaries
- USMLE and NCLEX-style MCQ preparation
- Medical image analysis (beta)
- Patient-facing condition explainers in simple language
AI-generated medical mock exams on demand. Free and no login required.
Includes:
- Realistic subject wise quesrions
- USMLE Step 1 style vignette exams
- USMLE Step 2 CK management-focused exams
- Timed testing environment
- Instant scoring and downloadable reports
Personal blog focused on:
- PICU and residency experiences
- Narrative medical writing
- Clinical cases
- AI in medicine
- Health explainers and medical technology
Python Kotlin Flask Jetpack Compose FAISS RAG LLMs On-device AI (LiteRT)
GitHub · Medium · PathToDL · PocketMed on Play Store
Building practical AI tools for real clinical problems.