const arpit = {
role: "Full-Stack AI Engineer & Hardware Developer",
base: "Nagpur, Maharashtra, India",
education: "B.Tech in Artificial Intelligence & Data Science — expected 2029",
operatingAt: [
"cloud-based Generative AI",
"secure backend architecture",
"real-time communication",
"responsive product engineering",
"ESP32 and embedded hardware"
],
engineeringMode: "production-first"
};I design systems where AI interfaces, web platforms, backend services, databases, real-time transport, and physical hardware operate as one product. My work spans animated Next.js frontends, Java Spring Boot and Node.js services, persistent authentication, WebSockets, WebRTC, MongoDB, Firebase, and ESP32 firmware.
The objective is not to make a demo that only works once. The objective is to build software that remains maintainable, observable, secure, responsive, and deployable when real users arrive.
|
Gemini-powered conversational systems, multimodal input, streaming responses, editable memory, personalization, and safe server-side orchestration. |
Java Spring Boot and Node.js services with validation, authentication, rate controls, health checks, structured failure handling, and database persistence. |
WebSocket/STOMP messaging, WebRTC signaling, presence events, uploads, session continuity, and responsive browser experiences. |
ESP32 and Arduino development, servo control, OLED interfaces, embedded C/C++, asynchronous loops, and cloud-to-device integration. |
|
Full-stack multimodal AI companion platform. Built around Gemini streaming with multi-conversation management, editable long-term memory, image/document attachments, speech, camera capture, reminders, persistent sessions, active-device controls, quotas, audit events, and an administrator surface.
|
Real-time private messaging and browser calling platform. Combines a React PWA with a Java Spring Boot backend, MongoDB-backed sessions, STOMP/SockJS messaging, GridFS media, Google OAuth 2.0, CSRF-aware cross-origin security, read receipts, chat pinning/muting, and WebRTC audio/video signaling.
|
|
Interactive engineering portfolio with an embedded AI assistant. A high-motion, responsive Next.js experience using React, TypeScript, Tailwind CSS, and a server-side Gemini route. It presents engineering work through a tech-noir interface, project navigation, contact flows, and an assistant grounded in portfolio information.
|
IoT and AI hardware integration. ESP32 firmware that coordinates servo motors, a 128×64 OLED interface, network-bound Gemini requests, and asynchronous hardware timing under real voltage, memory, and thermal constraints.
Domain: cloud AI ↔ embedded control ↔ physical interaction |
Note
Neosis uses encrypted HTTPS/WSS transport in production, but its stored text and media are not end-to-end encrypted. The backend can access persisted content.
01 Security is architecture, not a final middleware checkbox.
02 Real-time features require lifecycle control, backpressure, and failure recovery.
03 Authentication must survive browser restarts without exposing durable credentials.
04 UI animation must respect mobile performance and reduced-motion preferences.
05 AI output needs validation, limits, observability, and explicit operational boundaries.
06 Hardware code must account for timing, voltage, memory, thermals, and unreliable networks.
|
|
|
I am interested in engineering conversations and collaborations around AI platforms, production web systems, Java backends, real-time applications, developer tooling, and IoT hardware.
tagadearpit/tagadearpit.