Skip to content
View Noor-Sabah05's full-sized avatar
🎯
Focusing
🎯
Focusing

Block or report Noor-Sabah05

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
Noor-Sabah05/README.md

Noor ul Sabah Sami

CS @ FAST NUCES Lahore

πŸ“ Lahore, Pakistan

About Me

I build ML systems end-to-end β€” from raw feature engineering and model training to orchestrated pipelines, containerized deployments, and production-ready backends. My work sits at the intersection of applied machine learning, MLOps, and full-stack mobile development.

I'm particularly drawn to problems where the data is messy, the constraints are real, and the solution has to actually ship.

  • πŸ† Top 5 β€” CodeRush GDC ITU 2026 ML Competition (OOF Macro-F1: 0.741)
  • πŸ₯‰ 3rd Place β€” Build with AI Hackathon, GDGOC 2026
  • πŸŽ“ Currently pursuing BSc Computer Science (graduating Jun 2027)
  • πŸ“– AWS Academy Graduate β€” Cloud Foundations

πŸ† Competitions

Competition Result Stack
CodeRush GDC ITU 2026 ML Competition Top 5 LightGBM Β· XGBoost Β· CatBoost Β· Scikit-learn
Build with AI Hackathon β€” GDGOC 3rd Place Next.js Β· TypeScript Β· Turso Β· Gemini API

CodeRush: Engineered 174 features across 16,776 records via bag-level aggregation. Built a multi-seed ensemble with 5-fold CV across 25 runs, optimizing decision thresholds to hit OOF Macro-F1 of 0.741. Repo Link: CodeRush

Build with AI: Built SyncGuard in a timed hackathon β€” an AI decision-intelligence platform that detects conflicting commitments across startup teams using Gemini-powered conflict detection.


πŸ”¨ Featured Projects

TrendVerse β€” Multimodal Social Media Trend Forecasting PyTorch CLIP (ViT-B/32) HDBSCAN MLflow DVC Airflow Docker AWS Prophet

Replaced KMeans with CLIP embeddings + HDBSCAN for semantically coherent image clustering. Full Airflow-orchestrated pipeline, containerized with Docker, deployed on AWS, with Prophet-based virality forecasting.

ViseNotes β€” Cross-platform Voice Notes App Flutter FastAPI Speech-to-Text REST

Cut transcription time by 87.5% (5-min audio: ~20 min β†’ under 2.5 min) via async pipeline design and API batching.

Automer β€” Smart Android Automation Java Android Firebase

Context-aware automation with location-based mode switching, voice-triggered emergency activation, and intruder detection with Firebase cloud sync.

Job Recruitment System β€” Full-stack Hiring Platform React.js Node.js RESTful APIs

Role-based access for 3 user types across the full hiring lifecycle, with concurrent multi-role session management.


πŸ’» Tech Stack

ML / AI

Languages

Backend & Mobile

DevOps / Cloud

Databases

Pinned Loading

  1. DecodeLabs-Internship-data-science DecodeLabs-Internship-data-science Public

    Learning and projects done in decodelabs internship

    Jupyter Notebook 1

  2. trendverse trendverse Public

    TrendVerse is an AI-powered fashion trend forecasting and recommendation platform that predicts emerging trends using Prophet and delivers personalized recommendations using CLIP embeddings. Built …

    7

  3. vise_notes vise_notes Public

    Turns raw audio from lectures and meetings into structured, AI-powered notes. It transcribes speech, generates intelligent notes, extracts key insights, builds quizzes for learning reinforcement, a…

    Dart 7

  4. SyncGuard SyncGuard Public

    SyncGuard helps co-founders stay aligned by logging decisions, detecting conflicts with prior choices, and generating daily action briefings. It includes a decision log, conflict inbox, and history…

    TypeScript 7

  5. CodeRush-GDC-ITU-2026-Machine-Learning-Competition CodeRush-GDC-ITU-2026-Machine-Learning-Competition Public

    Economic class classification (lower/middle/upper) from bag-structured demographic data. Multi-seed 5-fold ensemble of LightGBM & XGBoost with OOF threshold tuning. Top 5 Β· Coderush 2026

    Jupyter Notebook 7