Skip to content
View Aashutosh31's full-sized avatar
🎯
Focusing
🎯
Focusing

Highlights

  • Pro

Block or report Aashutosh31

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
Aashutosh31/README.md
Typing SVG

┌──(aashutosh㉿arch)-[~/introduction]
└─$ cat about.md

🚀 2nd Year CS Student · Building distributed real-time applications and AI systems Creator of ARC-AI and WebChat — real-time messaging, voice/video infrastructure, and autonomous agent tooling, built end-to-end: architecture, backend, and deployment.

Portfolio LinkedIn

🧠 Featured Projects

┌──(aashutosh㉿arch)-[~/projects]
└─$ tree -L 1 --color=tokyonight

🌐 WebChat

Distributed Real-Time Communication Platform

Blue

A multi-paradigm social platform merging low-latency persistent messaging, Discord-style communities, Reddit-style forums, and ephemeral social feeds into one ecosystem.

Built solo: architecture, backend, real-time infrastructure, authentication, and deployment.

⚡ Engineering Highlights
  • Redis Pub/Sub keeps WebSocket message delivery synchronized across multiple server instances, so real-time events stay consistent as the backend scales horizontally
  • LiveKit powers multi-party WebRTC voice/video rooms, offloading media routing away from the application server
  • Payload encryption via the native Web Crypto API, paired with short-lived JWT access tokens and rotating refresh tokens, to limit exposure from a compromised session
  • Embedded ARC Assistant (Mistral AI + Pinecone vector memory) surfaces contextual support inside communities without a separate lookup step
  • Dockerized multi-container setup separates the API, real-time layer, and datastores into independently deployable services

React Node.js MongoDB Redis Socket.IO LiveKit Tailwind Docker

Repo Live

🤖 ARC-AI

Real-Time AI Assistant with Streaming Voice & Long-Term Memory

Violet

An AI agent that streams responses, executes tools, and maintains long-term memory — built on a full MERN-stack execution core.

Built solo: architecture, backend, agent runtime, RAG pipeline, and UI actuation.

⚡ Engineering Highlights
  • Token-by-token WebSocket streaming renders responses as they're generated instead of waiting for the full completion
  • Voice responses are interruptible — a buffered TTS pipeline stops playback mid-sentence once the user starts speaking
  • RAG-based memory retrieves relevant Pinecone embeddings from past conversations instead of replaying full history on every turn
  • Multi-step tool execution (web search, external APIs) runs without requiring a new prompt for each step
  • Recovery/retry logic sits above the LLM provider layer, so a failed call doesn't break tool_call continuity mid-execution

React Node.js WebSockets Vector DB TTS/STT RAG

Repo Live

🧩 Engineering Challenges

┌──(aashutosh㉿arch)-[~/challenges]
└─$ git log --oneline --grep="fix\|solve"
  • Synchronized real-time events across multiple server instances using Redis Pub/Sub, avoiding the single-instance WebSocket bottleneck
  • Built an interruptible TTS playback pipeline so voice responses can be cut off mid-sentence without breaking the buffering logic
  • Designed a semantic memory layer with Pinecone embeddings so the agent recalls relevant context instead of replaying full conversation history
  • Separated recovery/retry logic above the LLM provider layer in ARC-AI, preserving streaming continuity and tool_call consistency when a request fails
  • Containerized the full stack (API, real-time layer, datastores) with Docker for consistent local and production environments

⚙️ What I Work On

┌──(aashutosh㉿arch)-[~/skills]
└─$ fastfetch --focus
🧠 AI AgentsLLMs, RAG, semantic memory, tool calling
Real-TimeWebSockets, WebRTC, streaming architectures, low-latency sync
🛠 BackendNode.js, distributed system patterns, database optimization
📦 DevOpsGit workflows, Docker orchestration, CI/CD

🛠 Tech Stack


🎨 FrontendReact · Tailwind CSS · Framer Motion · Zustand
🧩 BackendNode.js · Express · Socket.IO
🗄️ Data & CacheMongoDB · Redis · Pinecone (Vector DB)
☁️ InfraDocker · LiveKit (WebRTC) · Vercel · Render
💻 Daily DriverArch Linux · Hyprland · Omarchy · VS Code

📌 Currently

+ Scaling WebChat infrastructure and expanding community features
+ Building and refining autonomous AI agents
+ Contributing to open-source workflows
+ Tuning the Hyprland rice, one pixel at a time

📊 Activity & Languages


📫 Reach Me

Portfolio LinkedIn GitHub


┌──(aashutosh㉿arch)-[~]
└─$ exit

Pinned Loading

  1. arc-ai-project arc-ai-project Public

    Realtime multimodal AI agent with voice streaming, RAG memory, and autonomous workflows

    JavaScript 2

  2. webchat-system-design webchat-system-design Public

    Architectural blueprints, system design, and API documentation for WebChat—a highly scalable, real-time social communication platform.

    2

  3. smart-attendance-backend smart-attendance-backend Public

    JavaScript 1

  4. dotfiles dotfiles Public

    Personal Arch Linux + Omarchy + Hyprland dotfiles with a production-ready development environment.

    Lua 1

  5. smart-attendance-frontend smart-attendance-frontend Public

    JavaScript 1

  6. Stock-Suggestion-Application Stock-Suggestion-Application Public

    JavaScript 1