M.Sc. Data Science student @ Hochschule Fulda 🇩🇪 | Former Data Analyst (3 yrs)
I build data pipelines, ML models, and BI dashboards. In my previous role as a Data Analyst, I cut manual reporting effort by 40%, sped up data processing by 30%, and freed 15+ hours/week of team capacity through Power Automate workflows — with zero reporting restatements over three years.
Currently seeking Werkstudent / internship roles in Data Science & Analytics in Germany.
- 🎓 M.Sc. Data Science, Hochschule Fulda (2025–present) · B.Sc. Computer Science
- 🌍 Fulda, Germany · English (C2) · German (A2, actively improving) · Urdu (native)
- 📫 Reach me: danishma309@gmail.com · LinkedIn
Languages & Querying
Machine Learning & Data
BI, Automation & Cloud
| Project | What it does | Stack |
|---|---|---|
| Fairness in ML - Paper Replication | Replicated bias-mitigation experiments from arXiv:2412.09896 — pre-, in-, and post-processing on Adult Income; post-processing cut demographic parity difference by 99.7% at a 2.2pp accuracy cost | Fairlearn, AIF360, XGBoost |
| Bilingual Tweet Classification | Semi-supervised classification of code-switched English/Roman-Urdu tweets into 6 categories, up to 92% accuracy — B.Sc. FYP with a co-authored research paper | scikit-learn, SVM, Naive Bayes |
| Vehicle Price Prediction | Benchmarked 5 regressors on 1M vehicle listings — gradient boosting and an entity-embedding neural net reach ~4–5% MAPE where linear regression fails at 33% | CatBoost, LightGBM, XGBoost, Keras |
| Crypto Market Dashboard | Real-time crypto analytics on a decoupled architecture: FastAPI backend shielding a Streamlit/Plotly frontend from the CoinGecko API | FastAPI, Streamlit, Plotly |
| ERA-5 Atmospheric Analysis | Exploring a 4-D global climate reanalysis grid (37 pressure levels, 0.25°) — memory-safe NetCDF handling, interpolation, multi-scale visualization | xarray, NetCDF, Matplotlib |
| Fashion-MNIST CNN Pipeline | End-to-end image classification as CLI tools — validation, preprocessing, and a Keras CNN with leakage-safe class balancing | TensorFlow/Keras, OpenCV |
| Mandelbrot Smooth Coloring | Research project: plugin-based coloring architecture + C0/C1 continuous iteration counts for artifact-free deep-zoom fractal rendering | NumPy, TensorFlow, plugin architecture |
| MNIST Digit Recognition on AWS | End-to-end cloud ML pipeline: training, serverless inference, storage, and load-balanced serving | EC2, Lambda, S3, DynamoDB, ELB |