Skip to content
View Ammar-Alnagar's full-sized avatar
🎰
Deciphering the GPU manuscript.....
🎰
Deciphering the GPU manuscript.....
31 followers · 10 following

Achievements

Achievement: Pull Sharkx3Achievement: Pair ExtraordinaireAchievement: StarstruckAchievement: YOLOAchievement: Quickdraw

Achievements

Achievement: Pull Sharkx3Achievement: Pair ExtraordinaireAchievement: StarstruckAchievement: YOLOAchievement: Quickdraw

Highlights

Organizations

@dev-protocol @cloudcommunity @Py-Contributors @Design-and-Code @recodehive @Coders-Evoke-Community @Jaidevstudio @WebXDAO @Bauddhik-Geeks @CommunityPro @Magic-Academy @Devs-Dungeon @infraform @yfosp @Th3-Shield @GrowInCommunity @Open-Hub-Community @Krypto-Hashers-Community

Block or report Ammar-Alnagar

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 supported. This note will be visible to only you.
Report abuse

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

Report abuse
Ammar-Alnagar/README.md

Hey, I'm Ammar Alnagar

Lead AI Systems Engineer | M.Sc., Artificial Intelligence

I translate cutting-edge research into production-grade AI infrastructure, specializing in ultra-low latency and high-throughput LLM deployment.

Core Focus

My work targets the bottleneck of generative AI: speed and stability at scale.

  • Accelerated Inference: Pushing model limits using low-level optimization (CUDA, Triton, Mojo).
  • System Engineering: Building robust systems for model serving, monitoring, and fault-tolerance.
  • Agent Orchestration: Designing fast, reliable multi-agent control flows.

Currently Working On

I am actively focused on performance optimization and serving frameworks:

  • vLLM: Maximizing GPU utilization via continuous batching and custom scheduling.
  • sglang: Building fast, reliable agents with structural generation and complex control flow.
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢀⣤⡶⠿⠿⠷⣶⣄⠀⠀⠀⠀⠀
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⣰⡿⠁⠀⠀⢀⣀⡀⠙⣷⡀⠀⠀⠀
⠀⠀⠀⡀⠀⠀⠀⠀⠀⢠⣿⠁⠀⠀⠀⠘⠿⠃⠀⢸⣿⣿⣿⣿
⠀⣠⡿⠛⢷⣦⡀⠀⠀⠈⣿⡄⠀⠀⠀⠀⠀⠀⠀⣸⣿⣿⣿⠟
⢰⡿⠁⠀⠀⠙⢿⣦⣤⣤⣼⣿⣄⠀⠀⠀⠀⠀⢴⡟⠛⠋⠁⠀
⣿⠇⠀⠀⠀⠀⠀⠉⠉⠉⠉⠉⠁⠀⠀⠀⠀⠀⠈⣿⡀⠀⠀⠀
⣿⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢹⡇⠀⠀⠀
⣿⡆⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⣼⡇⠀⠀⠀
⠸⣷⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢠⡿⠀⠀⠀⠀
⠀⠹⣷⣤⣀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⣀⣰⡿⠁⠀⠀⠀⠀
⠀⠀⠀⠉⠙⠛⠿⠶⣶⣶⣶⣶⣶⠶⠿⠟⠛⠉⠀⠀⠀⠀⠀⠀

Pinned Loading

  1. Helios-Engine Helios-Engine Public

    Helios Engine is a powerful and flexible Rust framework for building LLM-powered agents with tool support, chat capabilities, and easy configuration management. Create intelligent agents that can i…

    Rust 40 3

  2. Rust-TUI-Coder Rust-TUI-Coder Public

    A powerful terminal-based coding assistant that combines the convenience of a modern TUI with the intelligence of large language models. Rust TUI Coder provides an interactive environment where you…

    Rust 13 1

  3. Marla Marla Public

    This project implements an agentic pipeline using the Google Agent Development Kit (ADK) framework. It features a master agent that supervises and delegates tasks to a team of specialized agents, e…

    Python

  4. SLRAG-with-COT SLRAG-with-COT Public

    Self Learning RAG With COT is a project that implements Retrieval-Augmented Generation (RAG) combined with Chain of Thought (COT) reasoning. This project aims to enhance the performance of language…

    Python 17 3

  5. PlushieAI-PI PlushieAI-PI Public

    A Python-based voice assistant project that leverages Google's Gemini AI for interactive chat, complemented by Google Cloud Text-to-Speech (TTS) for announcements and Gemini's own audio for chat re…

    Python

  6. VisoLearn VisoLearn Public

    🏆 Award-Winning AI-Powered Educational Technology | Enterprise-Grade Analytics & Insights

    Python