Android Engineer · Startup Founder · Pune, India
that live
on the device.
2 years. 65,000 organic users. Three shipped apps. On-device LLMs, offline-first sync, and zero cloud dependencies. Co-founded Genzopia Solutions and owned every layer: product, architecture, monetization.
work without the cloud.
On-device AI, not cloud AI
Quantized LLMs running entirely on-device via JNI/NDK. Private, fast, and functional with zero signal. No API call. No latency. No data leaving the phone.
Offline-first, not online-first
Room + Retrofit sync designed to survive disconnection. Users in low-connectivity environments get the full experience — state reconciles silently when signal returns.
Owned architecture, not committee decisions
Co-founded and built Genzopia solo. Every product decision, architecture choice, and monetization experiment was mine to make and mine to live with.
Real users, not demo users
65,000+ organic installs. 4.9 ★ across 10+ countries. Production-grade crash cycles, ANR resolutions, and Profiler sessions — not simulated load tests.
Not just side projects.
Built and scaled Offline AI entirely solo. No paid acquisition. Product-market fit earned through iteration.
Custom Android launcher. OS-level app blocking, per-app time limits, Stripe integration. 10+ countries.
Kotlin, Jetpack Compose, MVVM/MVI, Coroutines/Flow, Room, Retrofit, WorkManager, JNI/NDK.
Offline AI · FocusGuard · GenzCrop · Device Doctor. One on Amazon Appstore. One on Play Store with 30K+ installs.
Every inference runs on-device. ONNX + quantized models + semantic search. The network permission is not requested.
Ideathon 1.0 & 2.0. University Project Competition. Competing, winning, shipping.
All production. All on the Play Store or Amazon Appstore.
Private AI assistant. Quantized LLMs on-device via custom JNI/NDK inference pipeline. Semantic search over local embeddings. No cloud calls. 65,000+ users.
View Project →Custom Android launcher with OS-level app blocking and per-app time limits. Full payment integration via Stripe. Built UI/UX from scratch. 4.9★ in 10+ countries.
View Project →Real-time hardware diagnostics for camera, microphone, and sensors. Published on Amazon Appstore.
View Project →offline AI possible.
// Offline AI — inference pipeline
model_format: .task, .gguf, .litertlm
inference_runtime: JNI/NDK — custom C++ bridge
embedding_search: on-device semantic vector index
network_permission: NOT REQUESTED
cloud_calls: 0
data_location: YOUR DEVICE
// Android stack
language: Kotlin
ui: Jetpack Compose
architecture: MVVM / MVI
async: Coroutines + Flow
local_db: Room
sync: Retrofit + WorkManager
ai_tools: MediaPipe · OpenCV · ONNX · Gemma · Qwenthat needs to work offline?