Skip to content
View KayTheCoder-101's full-sized avatar
3 followers · 7 following

Block or report KayTheCoder-101

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.

Content in all repositories owned by your account will be closed.
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
KayTheCoder-101/README.md

Um E Kulsoom

Machine Learning & MLOps — Building end-to-end pipelines: ingestion → training → tracking → deployment

LinkedIn Email University


Certifications

AWS Data Scientist GitHub Foundations


Featured Projects

Real-Time Streaming Data Pipeline

Solo-built pipeline streaming live BTC/USDT trades from the Binance WebSocket API through Kafka into Spark Structured Streaming, landing raw and 1-min OHLC data as Parquet in a MinIO data lake, served via Spark Thrift Server to live Superset dashboards. Diagnosed and fixed a race condition that could corrupt Structured Streaming's exactly-once metadata log.

Kafka Spark MinIO Airflow Superset Docker

Log Anomaly Detection Dashboard

End-to-end system generating, ingesting, and analyzing application logs in real time. A FastAPI + PostgreSQL ingestion layer feeds an Isolation Forest model that flags anomalies — traffic spikes, server errors, suspicious IPs — on a live Streamlit dashboard.

FastAPI PostgreSQL scikit-learn Streamlit Docker CI/CD

Real-Time Market Movement Prediction

End-to-end pipeline ingesting live data from Yahoo Finance, Reuters, Reddit, and Twitter to predict market direction and volatility using RNN, LSTM, and GRU models. Full MLOps stack with a FastAPI REST API deployed on AWS EC2.

RNN/LSTM/GRU DVC MLflow Airflow AWS EC2

Automatic Headline Generator (NLP)

Designed and benchmarked three architectures — Seq2Seq BiLSTM, BART, and a custom T5+Hybrid Attention — on CNN/DailyMail. Custom model reached 91% of BART's performance with 45% fewer parameters (62M vs 139M).

Seq2Seq BART T5 ROUGE

Pakistani Politician Image Classification

Multi-class facial recognition system classifying 16 public figures on 1,280+ images. Fine-tuned ResNet-50 and EfficientNet to 90%+ accuracy, versioned with DVC, tracked via MLflow, deployed on AWS EC2.

CNN ResNet-50 EfficientNet DVC MLflow AWS

Tech Stack

Languages Python SQL C++ C# JavaScript

ML & Data scikit-learn pandas NumPy NLP Computer Vision

MLOps & Infrastructure Docker AWS MLflow DVC Airflow Kafka Spark GitHub Actions

Databases & BI PostgreSQL MySQL Power BI Tableau


Open to entry-level roles in AI/ML Engineering · Data Engineering · Data Science

Pinned Loading

  1. log-anomaly-dashboard log-anomaly-dashboard Public

    Python 1

  2. Real-Time-Streaming-Data-Pipeline Real-Time-Streaming-Data-Pipeline Public

    Python 1

  3. moonlightaliza/headline-generation-nlp moonlightaliza/headline-generation-nlp Public

    A collaborative project for our 6th semester, Fundamentals of NLP course. The project aims to generate concise and relevant headings for an article automatically.

    Jupyter Notebook

  4. moonlightaliza/market-sentiment-prediction-mlops moonlightaliza/market-sentiment-prediction-mlops Public

    Python

  5. moonlightaliza/pakistani-politician-cnn-classifier moonlightaliza/pakistani-politician-cnn-classifier Public

    Python

  6. neurosymbolic-drug-checker neurosymbolic-drug-checker Public

    Python 2