Skip to content
View mushiraz's full-sized avatar
  • Meta
  • Toronto, Ontario, Canada
  • 08:18 (UTC -05:00)
  • LinkedIn in/mushiraz

Organizations

@InflareXYZ

Block or report mushiraz

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
mushiraz/README.md

LinkedIn Email GitHub


Building at the intersection of systems and intelligence

About

Software engineer at Meta based in Toronto. I build large-scale backend systems and explore the intersection of systems programming and machine learning — from GPU-accelerated video decoding and distributed infrastructure to applied ML, drones, and LLMs.


Featured Work

High-performance GPU video decoder used across AMD's compute ecosystem. Contributed core decoding features and optimizations.

C++ GPU Video

Production-ready video analysis system with local LLaVA-7B inference, optimized frame extraction, and 100% local processing.

Python LLaVA Ollama

Retrieval-augmented generation system for querying large-scale document collections with local LLM inference.

Python RAG LLM

Tello drone integration with OpenCV for real-time face detection and autonomous following capabilities.

Python OpenCV Drone


Tech Stack

LANGUAGES

Python C++ Rust TypeScript Java C# JavaScript

ML / DATA

PyTorch TensorFlow OpenCV scikit--learn Pandas

INFRASTRUCTURE

Docker Kubernetes Linux Jenkins MongoDB Azure

FRAMEWORKS

React Django Node.js Express Tailwind


Stats

  


Open to interesting conversations — feel free to reach out.

Pinned Loading

  1. ROCm/rocDecode ROCm/rocDecode Public

    [DEPRECATED] Moved to ROCm/rocm-systems repo

    C++ 33 25

  2. Face-Tracking-Drone Face-Tracking-Drone Public

    Utilizing Tello drone with OpenCV to do face detection and following.

    Python 2

  3. Gradient-Descent Gradient-Descent Public

    Implementation of the gradient descent algorithm which is commonly used for optimization in Machine Learning training algorithms.

    Python

  4. Language-Translator Language-Translator Public

    Allows for user to translate between English and German!

    Jupyter Notebook

  5. ollama_rag ollama_rag Public

    RAG based implementation with OLLAMA to serve as the API endpoint for querying large scale documents.

    Python

  6. video_inferencing_pipeline video_inferencing_pipeline Public

    🎬 Initial Release: Production-Ready Video AI Analysis Pipeline Complete enterprise-grade system with local Ollama/LLaVA-7B integration. Features: 100% local processing, optimized frame extraction, …

    Makefile