Santosh Subedi

Full-stack and AI product engineer. I design systems, write the decisions down, and build the whole thing: LLM and RAG features, native mobile apps, and the TypeScript services underneath.

Kathmandu, NepalFounding Engineer @ Rayu.aiConsultant CTO @ AIO Tech

Portrait of Santosh Subedi

Experience

  1. May 2026 – PresentFull-timeRemote

    Founding Engineer, AI & Product @ Rayu.ai

    Building Rayu from the ground up, leading AI, backend, and product engineering to create a personalized nutrition experience powered by wearable data.

  2. Dec 2025 – PresentFull-timeUnited Kingdom · Remote

    Software Engineer @ LaunchBox

    AI-powered full-stack web and mobile products across multiple client projects, including a video-streaming fitness platform.

    • Build LLM and RAG features: photo-based nutrition macro extraction, and learning assistants that handle video transcription, OCR, and PII redaction.
    • Author custom native modules, including a BLE wearable SDK bridge and an agentic notification engine with rule-based safety guardrails.
    • Responsible for frontend, backend, and the third-party integrations in between: payments, video, commerce, analytics.

  3. Sep 2023 – PresentConsulting · Part-timeKathmandu, Nepal

    Consultant CTO @ All In One Tech

    AIO Tech (aio.com.np) builds NFC-based solutions for organizational operations and acquired the All Attend attendance platform from BurgeonTech.

    • Guide the technical direction of all-attend, the attendance platform I originally architected at BurgeonTech.
    • Own the platform architecture and infrastructure, and set the roadmap as all-attend grows into more schools and enterprises.

  4. Apr 2023 – Dec 2025Full-timeAarhus, Denmark · Remote

    Full-Stack Developer @ Konsult-it

    Delivered 4+ production applications, all fully internationalized in Danish and English, for B2B clients across Europe.

    • Architected a real-time monitoring platform with MQTT/WebSocket communication, inventory tracking, and role-based access control.
    • Built a multi-tenant employee-benefits system: company admin portals, employee storefronts, and a seasonal gift shop with credit-based purchasing.
    • Developed payment processing with 3-D Secure, dynamic pricing rules, discount engines, and product configurators for 12+ print products.

  5. Sep 2020 – Aug 2023Full-timeKathmandu, Nepal · Hybrid

    Software Engineer @ BurgeonTech

    Real products used by schools and businesses across Nepal and abroad.

    • Designed the complete backend and infrastructure behind Nizi Tap attendance hardware with real-time syncing across locations (all-attend.com, later acquired by All In One Tech).
    • Built a school management system, an HR system backend, and AI chatbot agents with LLM integration and business-context automation.
    • Ran cloud deployments on AWS with Docker, and real-time messaging over MQTT and RabbitMQ.

  6. Jan 2021 – Apr 2023FreelanceRemote

    Full-Stack Developer @ Upwork

    Freelance full-stack web and mobile development for international clients, in parallel with full-time work.

Capabilities

AI-native products

AI as part of a real product, with deterministic boundaries around what it may do.

  • LLM integration
  • RAG & vector search
  • tool-use agents
  • voice
  • Vercel AI SDK
  • Gemini
  • Anthropic
  • OpenAI
Backend & distributed systems

APIs, queues, and real-time messaging that hold up when devices and users misbehave.

  • Node.js
  • Bun
  • Express
  • Hono
  • tRPC
  • MongoDB
  • PostgreSQL
  • Redis
  • BullMQ
  • RabbitMQ
  • MQTT
Product engineering

Web and mobile apps with the unglamorous parts done right: auth, payments, i18n, multi-tenancy.

  • React
  • Next.js
  • React Native / Expo
  • Vue 3
  • TypeScript
  • Stripe
  • Shopify
Infrastructure & operations

Shipping is half the feature; the other half is running it in production.

  • AWS
  • Docker
  • GitHub Actions
  • PM2
  • EAS Build
  • Terraform

About

I help founders and teams turn ideas into shipped products, on the web, on phones, and increasingly with AI inside. Over the last 6+ years I've done that for clients in the UK, Denmark, and Nepal. Some of that work runs on classroom RFID readers; some of it is an employee benefits store; one of them is an AI nutritionist.

I practice docs-first, architecture-first engineering. Before writing code I map the domain and put the important decisions in writing. Anyone who joins the project later can read why the system looks the way it does.

  • Understand the domain before writing code.
  • Write the important decisions down.
  • Keep systems simple until complexity earns its place.
  • Treat production as part of the feature.
  • Put deterministic boundaries around AI.

Get in touch

Thinking about bringing AI into your product? I help teams work out where LLMs actually fit, then build the integration and stay with it until it holds up in production. If you want a second opinion on an AI roadmap, or someone to ship it with you, write to me.

I'm also glad to talk about hard engineering problems and product architecture in general.

santsubcode@gmail.com