Building the operating layer for intelligent machines — from quantized local LLMs and real-time communication protocols to cross-platform AI code generation systems at CNEAPEE.
I'm Sanskriti Bhushan — a Full Stack Engineer at CNEAPEE Services, where I operate at the intersection of systems engineering and artificial intelligence. My current mandate is ambitious by design: to architect and ship CNEAPEE OS, a next-generation AI ecosystem that unifies web-based and locally-running language models under a single, coherent runtime.
My engineering vocabulary spans the full vertical stack — from LLM quantization research (squeezing 4-bit precision inference out of models that previously demanded cloud-scale compute) down to the React components and REST APIs users actually touch. I believe the best AI products are the ones where the seams between intelligence and interface are completely invisible.
Before CNEAPEE, I honed my design sensibility as a UI/UX Designer Intern at Unified Consultancy Services, shipping the Texora-AI interface, and deepened my mobile engineering fundamentals as an Android Developer Intern with Google for Developers. Those two poles — rigorous interface craft and systematic engineering discipline — define how I build everything.
I'm completing my B.Tech in Computer Science & Engineering at Galgotias University (Class of 2026), concurrently shipping production systems. The classroom and the codebase are the same place for me.
The AI core of CNEAPEE OS — a multi-model intelligence platform capable of routing user queries between cloud-hosted and quantized local LLMs in real time. Features a streaming chat interface, model-aware response rendering, real-time data analysis feeds, and a subscription billing system. Designed for multi-platform deployment across Web, Windows, and Android from a unified codebase.
A peer-to-peer real-time communication hub built on Next.js, WebRTC, and Coturn (STUN/TURN). Supports anonymous text chat, video calls, and voice calls with guest link generation. Engineered for sub-100ms latency handshakes and resilient connection traversal across NAT environments.
An AI-driven content management interface developed at Unified Consultancy Services. Purpose-built to make AI generation tools accessible to non-technical teams — featuring a structured component library, real-time content preview, and a streamlined editorial workflow that reduced time-to-publish by a significant margin.
A four-year programme in computer science and engineering — studied in parallel with active industry work. The curriculum spans algorithms, operating systems, computer networks, and database design. The most meaningful learning, however, happened in the intersection: applying distributed systems theory to real architectures at CNEAPEE while simultaneously completing the formal coursework that contextualises that practice.
I'm always open to conversations about ambitious engineering challenges, collaborations, or roles where the work genuinely matters. Say hello.