BS Data Science @ FCIT (7th semester) • Machine Learning • HTR Research (team) • UI/UX (learning)

I build practical ML projects and write down what works and what doesn’t.
I’m part of a team working on handwritten text recognition (HTR) for Notescape, where we benchmark on standard datasets and focus on improving accuracy (CER/WER) through better training, decoding, augmentation, and error analysis.

• Strengthening ML fundamentals: EDA → features → classic models → proper validation
• Studying deep learning (foundations, training routines, evaluation)
• Strengthening UI/UX across my projects so results are clear, accessible, and visually consistent
• Improving repo structure and READMEs across learning projects

• Plantly (team project): plant recognition + care tips
  Uses the plant.id API for identification and Wikipedia scraping for concise care guidance.
  My role: care-tips pipeline and API integration.
  TypeScript, React, Flask/FastAPI
  Repo: https://github.com/rabbiyabukhari/ML_project-plantly-smart-plant-care

• Google Play Store: analysis, clustering, rating prediction
  Notebook-driven pipeline: data cleaning, EDA, KMeans clustering, and baseline regression (RandomForest/GradientBoosting) with MAE/MSE/R² reporting.
  Python, pandas, scikit-learn, matplotlib
  Repo: https://github.com/rabbiyabukhari/Google-Play-Store-Apps-Comprehensive-Analysis-Clustering-and-Prediction

• Artificial Intelligence (classic search algorithms)
  BFS, DFS, IDDFS, Greedy Best-First, and A* with notes and lab material.
  Python
  Repo: https://github.com/rabbiyabukhari/Artificial_Intelligence

• Dynamic Programming (learning repo)
  Solutions to core DP problems (knapsack, LIS, coin change, edit distance, etc).
  Python
  Repo: https://github.com/rabbiyabukhari/Dynammic-Programming-Codes

• Notescape (team project): AI-assisted study workspace
  My role: CI/CD setup (GitHub Actions), file-upload pipeline, PDF chunking for downstream processing/search, merging teammates’ work (PR reviews), UI/UX design for study flows; ongoing HTR research to improve accuracy on existing datasets (CER/WER benchmarking).
  TypeScript, React
  Repo: https://github.com/NotescapeAi/Notescape

• SOFTEC’25 – Machine Learning (36-hour sprint) — 21/25 overall
  This rattled me. Instead of deciding ML wasn’t my cup of tea, I asked the better question: where does my practice actually lack, what don’t I know yet, and what should I do next?

• PuCON'25 - Palm Print | Auth (two rounds) — 4/24 in Round 2 (jury evaluation)
  I briefly reached 1st on the live board. In the last six hours I couldn’t work because I was also competing elsewhere. I also made preventable mistakes:

  • Trained on Colab T4 across two Google accounts, so run and file tracking suffered
  • Training error oscillated and accuracy stalled near ~96%, and under pressure I hard-coded a threshold/percent
  • Limited evolutionary search to ~30 generations while others ran ~100+
  • Didn’t stand up a minimal RNN baseline in time

Languages: Python, SQL, TypeScript
ML: scikit-learn, model selection/validation, basic tuning
Data: pandas, NumPy, matplotlib
Apps/Tools: FastAPI, Git, Docker, Linux, Jupyter/Colab
Research (HTR): benchmarking on existing datasets, CER/WER evaluation, decoding strategies, augmentation, error analysis
UI/UX: basic Figma; layout and typography for data apps
Ops: CI/CD with GitHub Actions; repo hygiene and PR reviews

BS Data Science, FCIT · 7th semester

A Levels, The Lahore Alma
Selected subjects: Computer Science, Mathematics, Art & Design

O Levels, Garrison Academy for Cambridge Studies
Selected subjects: Mathematics, Physics, Chemistry, Computer Science