┌──────────────────────────────────────────────────────────────────────────┐
│ PROFILE LOADED · AEROX RUNTIME · STATUS: ACTIVE │
│──────────────────────────────────────────────────────────────────────────│
│ operator → Jayaprakash Dey │
│ stack → Full-Stack · AI/ML · Native Android │
│ runtime → Python · TypeScript · Kotlin · SQL · JavaScript │
│ paradigm → Distributed Systems · Real-Time UX · Edge Intelligence │
└──────────────────────────────────────────────────────────────────────────┘
I don't build apps. I build systems — architectures that hold under pressure, scale without ceremony, and render at the threshold of hardware.
My engineering discipline spans three tightly coupled verticals: I design real-time, interactive full-stack software systems where every layer — from database schema to UI shader — is deliberate. I reach for Three.js via React Three Fiber when the interface demands spatial presence: immersive 3D scenes, procedural geometry, and WebGL-powered cinematic frontends that communicate information through environment, not just layout.
At the data layer, I don't just query — I model. I work with complex relational and vector datasets through raw SQL, designing normalized schemas that survive schema drift and serve analytics pipelines without bloating the response cycle. When the task crosses into intelligence, I reach into the Scikit-Learn ML toolkit — training, evaluating, and deploying custom classification and regression models with rigorous validation protocols.
For deployment infrastructure, I operate on a Firebase + Google AI Studio + Vercel triad: Firebase handles auth flows, Firestore's real-time event bus, and serverless function orchestration; Google AI Studio empowers my local LLM engineering workflows using Gemma, enabling on-device inference, prompt tuning, and RAG pipelines without cloud dependency; Vercel closes the loop on zero-downtime, edge-optimized delivery.
On mobile, I build native Android applications in Kotlin — threading-aware, Compose-driven UIs with Architecture Component compliance and lifecycle precision.
Every project under the AEROX umbrella is an assertion: that precision engineering and premium design are not mutually exclusive.
 Primary ML & Scripting |
 Full-Stack Type Safety |
 Runtime & Browser Logic |
 ML Modeling & Pipelines |
 Relational & Vector DB |
 Local LLM / Gemma Workflows |
 Auth · Firestore · Functions |
 SSR · Edge · App Router |
 Component Architecture |
 WebGL · R3F · Spatial UI |
 Native Android · Compose |
 Mobile Architecture |
 Edge Deployment · CI/CD |
| ⚡ AI / ML Systems | 🌐 Full-Stack Architecture | 📱 Native Android |
|---|---|---|
|
Custom model training via Scikit-Learn Local LLM inference with Gemma Prompt engineering & RAG pipelines Feature engineering on structured data Model evaluation & validation suites Google AI Studio workflow automation |
Next.js App Router & Edge functions Three.js / R3F spatial UI systems Real-time data via Firestore streams TypeScript-first component architecture REST & GraphQL API design Vercel edge-optimized deployment |
Kotlin coroutine-driven concurrency Jetpack Compose declarative UI MVVM + Clean Architecture compliance Android Architecture Components Lifecycle-aware state management Material 3 design system integration |
╔══════════════════════════════════════════════════════════════════╗
║ "Build precisely. Deploy confidently. Ship fast." ║
║ — AEROX Manifesto ║
╚══════════════════════════════════════════════════════════════════╝
