Shuhel K.

I build healthcare, wellness and fitness apps for web and mobile, and add AI to them. My telehealth platform has been live for over 2 years across 200+ hospitals, and the Kinobody fitness app I built has passed 100k downloads. I'm an AI product engineer with 8+ years of production experience. I'm part of Toptal's vetted network (top ~3%) and the founder of Inseed, a boutique product studio on Upwork. Between direct work and the studio, I've delivered healthcare, fitness, AI, mobile, and web products for VocalMD, Hazel Health, Nightingale Labs, Kinobody, Carrier, National Debt Relief, and YuMe, from seed-stage startups to Fortune 500 teams. Healthcare, wellness and fitness work I have shipped: - VocalMD: HIPAA-compliant telehealth app for iOS and Android, built from scratch as the sole mobile engineer. Doctor and patient sides, end-to-end encrypted chat, voice and video calls, EPIC FHIR integration, live on both app stores for over 2 years - Hazel Health: frontend engineer on a US school-based telehealth platform. Video visits, EHR charting with ICD-10 codes, medication orders, and the provider, patient and support portals - Nightingale Labs: remote patient monitoring dashboards and clinical charts, one React Native codebase for web, iOS and Android - Kinobody: fitness platform for a creator with a 1M+ following. I built the iOS and Android apps in React Native and the Next.js admin portal: video courses, personalized workout plans and trackers, subscriptions. 100k+ downloads - YuMe: gambling recovery app with an anonymous peer community, chat, coaching, and end-to-end encryption - Multi-tenant, role-based access control (RBAC) for regulated platforms AI work I have shipped: - Multi-agent LLM pipelines inside a live, end-to-end-encrypted app - RAG and semantic search: OpenAI embeddings, vector search, ranking - Real-time voice AI with Gemini Live, backed by long-term vector memory (pgvector) - AI integration with OpenAI, Anthropic Claude, and Google Gemini AI for health and fitness products is where I'm focused now: clinical note summarization, patient intake assistants, personalized coaching, and search or chat over FHIR data. Patient data only goes to model providers that will sign a BAA. Web and mobile, both in production: - Web: Next.js and React dashboards, Node and NestJS services, Postgres and Supabase - Mobile: React Native and Expo, managed and bare, through App Store and Play Store review - 1M+ download enterprise app (Carrier) - Offline-first architecture, push notifications, deep links, native modules, OTA updates - TypeScript-first, with unit, integration, and E2E coverage (Jest, Detox, Playwright) Stack I work in daily: React Native, Expo, TypeScript, React, Next.js Node.js, NestJS, GraphQL, REST APIs OpenAI, Claude, Gemini, RAG, pgvector, vector search Supabase, Firebase, PostgreSQL, AWS, Docker, Vercel Python where a job needs it: a Dagster scraping pipeline and an embeddings recommendation service Real impact: - Shipped a HIPAA-compliant telehealth product that has been live over 2 years - Migrated a 1M+ visitors/month site from WordPress to Next.js SSR, roughly 20x faster - Took an inherited AI marketplace from ~50% complete to live on both app stores - Cut crash rates and stabilized releases in legacy codebases without stalling the roadmap How I work: I care about your product outcome, not closing tickets. I flag risk early, I write TypeScript you can still read in a year, and if it is on my plate it ships. I work from Kathmandu with daily overlap with US and EU hours, and I send Loom walkthroughs as work progresses, so you always know where things stand. Building or extending a healthcare, wellness or fitness app, or adding AI to one? Send me a message. A real person reads it and replies within a few hours.