Muhammad S.
Data Annotator
Experience
Sep 2024 - Present
1 year 3 monthsData Annotator
Turing
- Wrote category-specific image prompts and scored 5,000+ model outputs using Apple’s RLHF rubric.
- Built and applied prompt sets; evaluated grounding, reasoning, and compliance on every response, following structured guidelines.
- Conducted jailbreak, bias, and prompt-injection tests; logged errors, inefficiencies, and edge conditions.
- Reviewed model transcriptions and intent alignment; produced clear outcome notes marked as success / partial / failure.
- Performed frame-level tagging of clicks, drags, and UI states; generated timestamped annotations to train action-sequence GUI agents through video annotation.
- Created LaTeX/Markdown ground-truth pairs for OCR models; QA’d every sample for loss-free parsing and consistency.
Prediction of AQI of Lahore City using Machine-Learning Models
- Compiled a 20-year dataset (2003–2022) from Copernicus CAMS; five pollutants plus temperature, 65,000+ rows.
- Bench-tested ARIMA, SARIMAX, STL-Decomposition, and XGBoost; evaluated with MSE and Dynamic-Time-Warping.
- Selected ARIMA (0,1,0) as best model, reducing forecast error by 30% vs. alternatives; produced structured performance summaries.
Languages
Urdu
NativeEnglish
AdvancedEducation
Oct 2020 - Jun 2024
National University of Sciences and Technology
Bachelors in Engineering · Engineering · Islamabad, Pakistan
Certifications & licenses
Machine Learning Specialization
Stanford
Mathematics for Machine Learning
Imperial College London
Need a freelancer? Find your match in seconds.
Try FRATCH GPT More actions
Similar Freelancers
Discover other experts with similar qualifications and experience
- English
- Deutsch
2025 © FRATCH.IO GmbH. All rights reserved.