Mohaned Kamal
Data Scientist
Python
SQL
MATLAB
ABAP
Data Pipeline & Feature Engineering: 4/10
Active 2 days ago
Download CVCV
Overview
Technical skills
Projects
Timeline
Roles
Overview
Applied ML engineer (middle level) focused on end-to-end supervised learning for tabular problems and simple desktop prototyping. The strongest proven skill is building and comparing ensemble and tree-based models for tabular prediction tasks as shown by the end-to-end pipelines and model experiments in Rent_It/Rent_it.ipynb and Segmint_It/segment it.ipynb. There is little evidence of production-grade MLOps, experiment tracking, custom model architectures, or rigorous software engineering practices in the public code.
Phone
Technical skills
Languages
4
Python
SQL
MATLAB
ABAP
Python
3
Beautiful Soup
pySpark
Tkinter
DevOps
4
Azure DevOps
Git
Rest API
Azure
AI/ML
6
Airflow
NumPy
Pandas
Scikit-learn
XGBoost
CatBoost
Other
15
Selenium
Power BI
MS SQL
Databricks
Azure Data Factory
CDS Views
Computer Vision Toolbox
GitHub
Spark
Jupyter Notebook
CI/CD
AI/ML
Classic ML
NLP
ETL/ELT
Projects
Customer Segmentation & NLP Projects
Built Python-based customer segmentation workflows using preprocessing and clustering approaches. Applied K-Means and similarity measures (Jaccard distance) and evaluated clustering quality with metrics. Included tweet clustering experiments to support segmentation use cases.
Enterprise BI & Operational Data Modeling
Curated SQL-based reporting datasets feeding Power BI dashboards. Modeled and prepared data for retail sales, inventory, profitability, and supply chain KPIs. Validated outputs with stakeholders to ensure the dashboards reflect trusted operational metrics.
Retail Market Price Comparison & Data Collection
Created an automated pipeline to collect and normalize retail market pricing data. Used Python with web scraping tools to extract product and price inputs, then standardized them for analysis and comparison. The result supports structured pricing comparison and management reporting.
Enterprise Data Pipelines & ERP Integration
Implemented SQL Server transformations and Python-based data preparation for ERP operational datasets. Designed orchestration using Azure Data Factory and Apache Airflow, with Databricks/PySpark processing for transformation workloads. Added API-based data exchange for Dynamics 365 F&O and focused on reliable, validated datasets for reporting.
Timeline
Data Scientist
·
Middle
GSO – Ministry of Defence
·
Full-Time
Built and maintained SQL Server data extraction and transformation logic using T-SQL, reusable datasets, and reporting views. Orchestrated ingestion and transformations with Azure Data Factory and Apache Airflow, and processed data in Azure Databricks using PySpark and Python. Integrated Microsoft Dynamics 365 Finance & Operations data via REST APIs, performed cleaning and reconciliation, and implemented dataset validation for downstream reporting. Developed Power BI dashboards and automated retail data collection using BeautifulSoup and Selenium.
MS SQL
SQL
Azure Data Factory
Airflow
Databricks
pySpark
Python
Rest API
Beautiful Soup
Selenium
Power BI
Git
Azure DevOps
Data Scientist Intern
·
Junior
Saudi German Hospital
·
Internship
Developed SAP ABAP/Open SQL extracts and reporting outputs to retrieve, transform, and present operational business data. Worked with SAP Data Dictionary concepts, CDS Views, and OData services to understand reusable data models and integration points. Supported extraction and preprocessing with Python and SQL, including data validation and reporting assistance.
ABAP
SQL
CDS Views
Python
Ain Shams University
Bachelor's Degree ·
Computer Science
Middle AI/ML Engineer
Confidence: Medium ML Engineer
Applied ML engineer (middle level) focused on end-to-end supervised learning for tabular problems and simple desktop prototyping. The strongest proven skill is building and comparing ensemble and tree-based models for tabular prediction tasks as shown by the end-to-end pipelines and model experiments in Rent_It/Rent_it.ipynb and Segmint_It/segment it.ipynb. There is little evidence of production-grade MLOps, experiment tracking, custom model architectures, or rigorous software engineering practices in the public code.
Model Architecture & Training
3/10
How well models are designed and trained
Solid applied use of standard supervised learning models and ensembles (RandomForest, XGBoost, CatBoost, stacking). No evidence of custom architectures, custom loss functions, or training from-scratch research-grade model design.
Evidence
Rent_It/Rent_it.ipynb: use of RandomForestRegressor, GradientBoostingRegressor, CatBoostRegressor, XGBRegressor
Segmint_It/segment it.ipynb: CatBoostClassifier and StackingClassifier usage and model.fit/predict calls
Data Pipeline & Feature Engineering
4/10
How data is prepared for models
Practical tabular data preprocessing and feature engineering (date parsing, one-hot encoding, scaling, missing-value handling, sampling to rebalance). Pipelines are notebook-style and not productionized.
Evidence
Rent_It/Rent_it.ipynb: date parsing, month/day extraction, one_hot_encoding function, StandardScaler snippets
Pre-Cancer_Cell detection/pre_cancer_detect.py: heavy column dropping, replace '?', dropna, filtering by Codon, resampling to balance classes
Segmint_It/segment it.ipynb: null imputation, LabelEncoder, StandardScaler and MinMaxScaler uses
Experimentation & Evaluation
3/10
How results are measured and tested
Basic experimentation and evaluation practices present (train/test splits, cross_val_score, confusion matrices). Missing systematic experiment tracking, baselines comparison, or rigorous metric analysis beyond accuracy/MSE.
Evidence
Rent_It/Rent_it.ipynb: train_test_split, cross_val_score, mean_squared_error reported
Pre-Cancer_Cell detection/pre_cancer_detect.py: DecisionTreeClassifier training, cross_val_score and confusion matrix plotting
Segmint_It/segment it.ipynb: cross_val_score, confusion matrix plotting and model score prints
MLOps & Deployment
2/10
How models are shipped to production
Lightweight deployment artifacts: competition submission CSV generation and a Tkinter GUI for local interaction. No evidence of model versioning, API serving, monitoring, or CI/CD.
Evidence
Rent_It/Rent_it.ipynb: submission.to_csv(...) producing competition submission files
Pre-Cancer_Cell detection/pre_processing.py and detection.py: Tkinter GUI that calls the trained classifier for local predictions
Computational Efficiency
1/10
How efficiently computing resources are used
Minimal evidence of computational optimization; models are used with default settings and without profiling, batching, or GPU/parallelization engineering.
Evidence
Rent_It/Rent_it.ipynb: RandomForestRegressor(n_estimators=100) and CatBoostRegressor(iterations=1199) shown but without explicit efficiency/profiling code
Segmint_It/segment it.ipynb: standard .fit() calls for ensemble models without optimization traces
Research Depth & Innovation
1/10
Depth of research and new ideas
No original research contributions or custom model implementations; work is applied reproduction of known algorithms and standard pipelines.
Evidence
Pre-Cancer_Cell detection/pre_cancer_detect.py: domain-aware data filtering (Codon 248 and 249) and conventional classifier application
Segmint_It/segment it.ipynb: exploration of ensemble/stacking approaches but no novel algorithmic development
Expertise
Medical AI & Healthcare· Middle
Marketing & AdTech· Middle
Industries
Health Care· Middle
Sales & Marketing· Middle
Transportation & Logistics· Middle
Technologies
AI/ML
SQL· since 2024 · Junior
CatBoost
Spark
Airflow· since 2024
Databricks· since 2024
Jupyter Notebook
XGBoost
Azure DevOps· since 2024
Scikit-learn
Beautiful Soup· since 2024
NLP
Azure
MS SQL· since 2024
CI/CD
Pandas
NumPy
Git· since 2024
pySpark· since 2024
GitHub
Recommendations
- Develop competition-style or PoC tabular ML models and deliverables (feature engineering, ensemble tuning, submission pipelines).
- Build small productionization projects that add model serving and basic monitoring (wrap an existing trained model in a lightweight API and add simple tests).
- Create reproducible experiment notebooks run via a lightweight runner (add logging, parameterized configs, and an experiment-tracking tool like MLflow).
- Expand applied bioinformatics pipelines into modular scripts with unit tests and documented data contracts to increase reliability for healthcare use cases.
Repositories
The developer's experience in this domain has been verified based on AI analysis of the following repositories:
Middle Data Scientist
Confidence: Low ML Practitioner
A middle-level ML practitioner focused on applied computer vision and basic unsupervised NLP workflows. The strongest proven skill is hands-on image preprocessing and geometric alignment for QR-like markers, demonstrated by the MATLAB functions Main_Fun.m, Alignment.m and Multi_Qrs.m. There is little evidence of production hardening, automated testing, reproducible environments or rigorous statistical validation in the public code.
Statistical Rigor
2/10
Correct use of statistics
Minimal statistical rigor: clustering uses SSE only, no hypothesis tests, no cross-validation or uncertainty quantification.
Evidence
K-Mean_Tweets_Clustring/Tweets.py: k_means and SSE functions
K-Mean_Tweets_Clustring/Tweets.py: no cross-validation or validation logic around clustering results
Data Wrangling & Cleaning
4/10
Preparing and cleaning data
Basic data wrangling present: text preprocessing with URL/punctuation removal and encoding fallback; image preprocessing includes median filtering and local thresholding, but cleaning is ad-hoc and lacks provenance or automated pipelines.
Evidence
K-Mean_Tweets_Clustring/Tweets.py: preprocessing function and try-except file open with encoding fallback
Automatic-Segmentation-and-Alignment-of-QR-Codes/Automatic-Segmentation-and-Alignment-of-QR-Codes-main/Main_Fun.m: rgb2gray, medfilt2 and local block thresholding
Exploratory Analysis & Visualization
2/10
Exploring and visualizing data
Very limited exploratory analysis: MATLAB code includes debug visualizations and plotted centers, but there is no structured EDA, written interpretation, or storytelling.
Evidence
Automatic-Segmentation-and-Alignment-of-QR-Codes/Automatic-Segmentation-and-Alignment-of-QR-Codes-main/Main_Fun.m: figure/imshow and plot calls
Predictive Modeling
2/10
Building models that predict
Basic predictive/unsupervised modeling implemented (custom k-means with Jaccard distance) but lacks evaluation rigor, baselines, cross-validation, hyperparameter tuning or error analysis.
Evidence
K-Mean_Tweets_Clustring/Tweets.py: k_means and Jaccard_distance implementations
K-Mean_Tweets_Clustring/Tweets.py: SSE used as sole model metric
Business Insight & Impact
1/10
Turning analysis into business value
No business framing or impact analysis; outputs are printed but there is no mapping to business metrics, cost-of-errors, or actionable recommendations.
Evidence
K-Mean_Tweets_Clustring/Tweets.py: main loop prints cluster sizes and SSE without linking to downstream decisions
Reproducibility & Notebook Hygiene
1/10
Clean, repeatable analysis
Low reproducibility and notebook hygiene: standalone scripts, no pinned environment, no seeding, and reliance on interactive input; limited packaging or automation.
Evidence
K-Mean_Tweets_Clustring/Tweets.py: use of random.sample without a fixed seed
K-Mean_Tweets_Clustring/Tweets.py: no requirements.txt or environment specification in provided files
Industries
Health Care· Middle
Technologies
Classic ML
MATLAB· since 2024 · Middle
Computer Vision Toolbox
Machine Learning· mentioned only
Python· mentioned only
Recommendations
- Develop small-to-medium prototyping work: image preprocessing and geometric-alignment CV pipelines (QR/barcode detection) where quick iterations and visual debugging are required.
- Build experimental unsupervised-text workflows and clustering prototypes that require custom distance metrics and ad-hoc preprocessing.
- Invest time converting scripts into reproducible modules: add requirements/environment files, seeds, unit tests and a simple CI pipeline to enable safer iteration and collaboration.
- Strengthen model evaluation and data-science rigor: add baselines, cross-validation where applicable, error analysis and written interpretation of findings.
Repositories
The developer's experience in this domain has been verified based on AI analysis of the following repositories:
Junior Backend Developer
Confidence: Medium Generalist
A junior-level Python developer focused on small ML and desktop data tools with a practical inclination toward prototyping. The strongest proven skill is data preprocessing and simple model training as implemented in pre_cancer_detect.py (pandas-based cleaning, LabelEncoder, train/test split and a DecisionTreeClassifier training pipeline). There is little evidence of production backend concerns such as APIs, database migrations, robust error handling, structured logging, or scalability strategies.
API Design
How well APIs are designed
Not evidenced in public code
Data Layer & Database
2/10
Working with databases
Basic data-layer work using CSVs and pandas with ad-hoc preprocessing and label encoding; no database migrations, transactions, or tuned SQL are present.
Evidence
Pre-cancer_Cell-detection/Pre-Cancer_Cell detection/pre_cancer_detect.py: uses pd.read_csv and Dataframe.to_csv for I/O and pandas-based preprocessing
Pre-cancer_Cell-detection/Pre-Cancer_Cell detection/pre_cancer_detect.py: LabelEncoder usage and Dataframe dtype conversions
Scalability & Performance
1/10
Handling load and speed
Single-threaded scripts and algorithms with no explicit scalability, caching, queuing, or connection-pooling patterns; basic iterative k-means implementation without performance tuning.
Evidence
K-Mean_Tweets_Clustring/Tweets.py: k_means, assign_clusters and SSE implement in-memory clustering without optimization
Pre-cancer_Cell-detection/Pre-Cancer_Cell detection/pre_cancer_detect.py: sampling and train_test_split but no profiling or scaling strategy
System Architecture
2/10
Overall system structure
Simple module separation (UI modules vs model/processing module) showing basic separation of concerns for a desktop app; no service decomposition, configuration management, or inter-service contracts.
Evidence
Pre-cancer_Cell-detection/Pre-Cancer_Cell detection/pre_processing.py, detection.py, home_page.py: distinct UI modules using tkinter
Pre-cancer_Cell-detection/Pre-Cancer_Cell detection/pre_cancer_detect.py: separated data processing and model training logic
Security & Auth
1/10
Protecting data and access
Minimal or no security controls: no authentication, authorization, input sanitization beyond basic CSV reading, and no secrets handling; potential unsafe direct file uses.
Evidence
Pre-cancer_Cell-detection/Pre-Cancer_Cell detection/detection.py: direct use of filedialog result passed to model test without validation
Pre-cancer_Cell-detection/Pre-Cancer_Cell detection/pre_cancer_detect.py: direct pd.read_csv usage with hardcoded or user-supplied paths and no validation
Reliability & Observability
2/10
Stability and monitoring
Some basic error handling and hygiene (warnings suppression and a UnicodeDecodeError fallback), but no structured logging, metrics, timeouts, retries or graceful shutdown patterns.
Evidence
Pre-cancer_Cell-detection/Pre-Cancer_Cell detection/pre_cancer_detect.py: warnings.filterwarnings('ignore') present
K-Mean_Tweets_Clustring/Tweets.py: fallback try/except for UnicodeDecodeError when opening files
Expertise
Python· Junior
Industries
Health Care· Middle
Data & Analytics· Junior
Technologies
Python· since 2024 · Middle
Rest API· since 2024
Tkinter
Java· mentioned only
Text· mentioned only
Recommendations
- Develop small-to-medium ML prototypes and data-cleaning pipelines where pandas and scikit-learn are primary tools.
- Build single-user desktop utilities or internal tools with Tkinter for rapid data exploration and demos.
- Work on improving robustness: add structured logging, input validation, and unit tests around preprocessing and model code.
- Contribute to backend-focused projects to gain experience with APIs, DB migrations, and production-grade reliability patterns (timeouts, retries, graceful shutdown).
Repositories
The developer's experience in this domain has been verified based on AI analysis of the following repositories: