Skip to content
View samarthsaxena2004's full-sized avatar
:atom:
welcome, sneaking my profile!?
:atom:
welcome, sneaking my profile!?

Block or report samarthsaxena2004

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
samarthsaxena2004/README.md

I'm Samarth from India. I build AI applications, agentic workflows, distributed backends, and low-latency systems. I enjoy experimenting with new ideas and building practical, scalable systems, with a focus on LLM inference, systems engineering, and fault-tolerant architectures.

portfolio  •  twitter  •  linkedin  •  mail

Projects

Here are some of my highlighted projects:

Context-aware desktop assistant that sees what you see on screen with sub-second response streaming and voice transcription.

  • Tech Stack: TypeScript, Electron, React, Cerebras GLM-4.7, Deepgram, Turborepo
  • Recognition: 1st Place Winner in the Cerebras GLM-4.7 Hackathon ($2,500 USD Prize).

Multi-agent document intelligence platform digitizing unconstrained handwritten forms into structured JSON in ~12 seconds.

  • Tech Stack: TypeScript, Next.js, Cerebras, Gemma 4 (31B), TailwindCSS
  • Key Highlight: 5-agent pipeline handling visual field extraction, handwriting recognition, and schema validation.

Self-healing distributed task queue and worker pool system with autonomous error mitigation.

  • Tech Stack: Go (Golang), Redis, Google Gemini Pro, Docker
  • Key Highlight: Intercepts dead-letter queue (DLQ) task crashes and leverages Gemini Pro to diagnose stack traces and apply runtime remediation.

AI research copilot converting complex PDFs into interactive knowledge graphs, 3D concept flashcards, and audio discussions.

  • Tech Stack: TypeScript, Next.js 16, RAG, LangChain, Vector Databases, Web Audio
  • Key Highlight: Hybrid semantic vector search paired with audio-synchronized concept breakdowns.

AI-native procedural roguelike driven by an autonomous Dungeon Master agent and real-time generative UI.

  • Tech Stack: TypeScript, React, Tambo Agentic Intelligence, TailwindCSS
  • Key Highlight: Dynamic game loop that generates branching narratives and on-the-fly UI tailored to player choices.

Header-only, high-performance C++ thread pool with a thread-safe task queue and asynchronous return value retrieval.

  • Tech Stack: C++17/20, Multithreading, Concurrency Primitives (std::future, std::condition_variable)

Check out my full repositories for more.

Pinned Loading

  1. note-wave note-wave Public

    Tired of boring textbooks? NoteWave uses RAG to turn any PDF into research chats, 3D flashcards, and audio-synced AI host conversations.

    TypeScript 1

  2. PlayDungeon PlayDungeon Public

    AI-native roguelike powered by Tambo Agentic Intelligence. Features generative UI, voice-driven gameplay, and a Dungeon Master.

    TypeScript 1

  3. flickAI flickAI Public

    Intelligent, context-aware desktop assistant that sees what you see. Powered by Cerebras GLM 4.7 and Cline.

    TypeScript 2 1

  4. prism prism Public

    Document Intelligence by Cerebras x Gemma 4 — 2026

    TypeScript 2

  5. resq-orchestrator resq-orchestrator Public

    A Self-Healing Distributed Job Queue System powered by Go, Redis, and Gemini.

    Go 1

  6. modern-cpp-mastery-lib modern-cpp-mastery-lib Public

    Tried creating my own library in C++

    C++