Skip to content
View SHREY-FR4NKL1NNN's full-sized avatar

Highlights

  • Pro

Block or report SHREY-FR4NKL1NNN

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
SHREY-FR4NKL1NNN/README.md

Typing SVG


LinkedIn GitHub


About

BTech CSE (AI/ML) · Polaris School of Technology

Builder focused on local AI infrastructure, agentic systems, and full-stack development. I run LLMs locally on consumer hardware and build production-quality systems on top of them — from MoE routing layers to real-time neural network observatories.


Featured Projects

MoE-inspired local LLM routing layer

Decomposes queries into sub-tasks, routes each to the right expert (Mistral 7B · DeepSeek R1 · LLaVA · Llama 3.2), runs them in parallel via asyncio, streams responses via SSE.

Python FastAPI React PyTorch

CI Live Demo

Real-time neural network training observatory

GPU-backed (RTX 4060, CUDA 12.6) training with live weight heatmaps, gradient norms, and activation histograms streaming via SSE. Memory-augmented input via text embeddings.

Python PyTorch CUDA React

CI Live Demo


Stack

Python PyTorch CUDA FastAPI React Next.js TypeScript Docker


GitHub Stats


Experience

Tectonic (×2 internships, Bangalore)

  • Frontend Engineer — e-commerce storefronts for 15+ brands (React, Next.js, WebSockets)
  • Shopify Developer — Spectrum AI app (Liquid, AI integration, MCP workflows)

Profile Views


Pinned Loading

  1. localmind localmind Public

    MoE-inspired local LLM routing — Mistral 7B + DeepSeek R1 + LLaVA + Llama 3.2, asyncio parallel execution, SSE streaming, rule-based gating network, fully local

    Python

  2. neuralcanvas neuralcanvas Public

    Real-time neural network training observatory — GPU-backed PyTorch, live weight heatmaps + gradient norms + activation histograms via SSE, memory-augmented input with text embeddings

    Python

  3. taskboard-pro taskboard-pro Public

    Real-time Kanban task board backend — Express + MongoDB/Mongoose, Socket.io live updates, JWT + Firebase auth, email notifications and cron reminders

    JavaScript

  4. dotfiles dotfiles Public

    Dev environment config for Windows 11 + WSL2 — bash aliases/PATH, Git identity, VS Code settings, and WSL2 resource limits

    Shell