Open to opportunities

Android Developer
building with AI

I'm Santhosh J — a generalist engineer with deep roots in Android development, now expanding into generative AI, IoT systems, and developer tooling. I ship production-grade code across mobile, web, and infrastructure.

30+
Public Repositories
9+
Years of Experience
8+
Languages Used
2025
AI Transition Year

From Android to AI

I started my career building Android apps — Jetpack Compose, MVVM, Clean Architecture, the whole stack. I've shipped production apps, published open-source libraries, and built Android samples that demonstrate best practices.

But I've always been a generalist at heart. Over the last year, I've been teaching myself generative AI engineering — building with LLMs, RAG pipelines, voice AI, multi-agent systems, and developer tooling powered by Claude Code.

Today I work across the full stack: Go for performance-critical services, Python for AI backends, TypeScript/Next.js for web apps, Kotlin for Android, and everything in between. I believe the best engineers are the ones who can move fluidly between domains.

  • Android Foundation

    9+ years of Kotlin, Jetpack Compose, MVVM, Clean Architecture, Coroutines, Hilt, Room, Paging3

  • Full-Stack Expansion

    Built IoT simulators in Go and Python, web dashboards in SvelteKit and React, serverless backends with Supabase

  • AI Engineering (2025–Present)

    LLM integration, RAG, voice AI, multi-agent systems, Claude Code plugins, semantic search with local embeddings

  • Open Source & Tooling

    Published Android libraries, Claude Code plugins, developer tools used by the community

Tech Stack

A broad toolkit built across mobile, AI, web, and infrastructure — always learning, always shipping.

🤖

Generative AI

LLM Integration RAG Pipelines Voice AI Multi-Agent Systems Semantic Search Vertex AI
📱

Android / Kotlin

Jetpack Compose MVVM Clean Architecture Coroutines Hilt Room Paging3
🌐

Web Development

Next.js React SvelteKit TypeScript Tailwind CSS Fastify
⚙️

Backend & APIs

Go Python FastAPI Supabase Deno PostgreSQL
🔧

Infrastructure & DevOps

Docker GitHub Actions GHCR MQTT Home Assistant Nginx
🧪

Testing & Quality

pytest JUnit CI/CD Load Testing Smoke Tests

What I've Built

Selected projects from the last 6 months — spanning AI tools, IoT platforms, Android libraries, and developer tooling.

🎧

VoiceRead

A local-first PDF listening app — upload any PDF and listen with natural AI voices powered by Sarvam AI TTS. Real-time sentence highlighting, ~40 Indian English voices, smart audio caching. A self-hosted alternative to Speechify.

JavaScript Next.js + Fastify + SQLite
🧠

Engram

Semantic memory for Claude Code — find any past session by meaning, not keywords. Local-first embeddings with ONNX, SQLite vector search, automatic session indexing via hooks. Search the way you think.

JavaScript ONNX + SQLite-vec
🤝

TeamWright

Claude Code plugin that plans, designs, creates, and launches complete multi-agent teams from a single goal. 12-step orchestration pipeline with 3-layer validation and resilient recovery.

Shell Claude Code Plugin
IoT
View →
🏠

Nodrithm Dev Simulator

Production-grade IoT simulator for smart hostel management. Simulates 10,000+ virtual smart switches with realistic energy patterns, Home Assistant auto-discovery, real-time SvelteKit dashboard, and Go backend. Single Docker image, multi-arch support.

Go SvelteKit + MQTT
IoT
View →

Nodrithm PowerHub

IoT device simulator for testing smart home automation. Emulates Shelly 1PM Mini Gen4 switches with full RPC-based MQTT, Home Assistant discovery, pytest suite (80% coverage), and React dashboard. 6+ months of iterative development.

Python FastAPI + React
Android
View →
🎤

VoiceKeyboard

Android app with a persistent floating bubble for speech-to-text transcription from anywhere on the device. Powered by Moonshine STT with on-device model loading, accessibility service integration, and modular MVVM Clean Architecture.

Kotlin MVVM + 4 Modules
Android
View →
🛡️

Screenshot Blocker

Production-ready Android library that prevents screenshots and screen recording using FLAG_SECURE. Zero dependencies, JitPack-published, used in production apps. Supports API 21+ with one-line initialization.

Kotlin JitPack · v1.3.0
Tool
View →
🧹

macOS Storage Cleanup

Claude Code plugin for comprehensive macOS storage analysis — scans Xcode DerivedData, Android Studio caches, Docker images, iOS backups, and more. Safe trash-based cleanup with two-tier safety and dry-run by default.

Shell Claude Code Plugin
🏥

Clearscope

AI-powered medical imaging analysis using Google's MedGemma multimodal LLM via Vertex AI. Clean Architecture with Hexagonal patterns, FastAPI backend, Next.js frontend, PostgreSQL. Upload X-rays, get AI analysis.

Python FastAPI + Next.js + Vertex AI
Tool
View →
🏨

Bedrock HMS

Hostel management platform built on Supabase Edge Functions with TypeScript/Deno and PostgreSQL Row-Level Security. 22 edge functions covering full CRUD with role-based access control and automatic audit trails.

TypeScript Supabase + Deno + RLS

Where I've Worked

My professional journey from Android engineer to a generalist building across the full stack.

Senior Android Engineer

Nodrithm
2022 – Present
  • Built and maintained production Android apps using Jetpack Compose, MVVM Clean Architecture, and Hilt dependency injection
  • Designed and implemented IoT device simulation platforms in Go and Python for testing smart home automation at scale
  • Architected full-stack solutions including MQTT infrastructure, Home Assistant integrations, and real-time web dashboards
  • Published open-source Android libraries (Screenshot Blocker) used by the developer community
  • Led the transition from legacy Python simulator to a high-performance Go-based platform with 10,000+ device capacity

Generative AI Engineer (Self-Directed)

Independent Projects
2025 – Present
  • Built AI-powered applications integrating LLMs, voice synthesis (Sarvam AI), and multimodal models (MedGemma via Vertex AI)
  • Developed Claude Code plugins for multi-agent orchestration (TeamWright), semantic memory (Engram), and system tooling
  • Implemented local-first AI features: ONNX embeddings, SQLite vector search, semantic session recall
  • Designed RAG pipelines, voice AI agents, and multi-agent systems with validation and recovery patterns
  • Created developer tools and plugins used by the AI engineering community