Expert Setup for Machine Learning Research Environment

I need an expert to fully set up my research environment on my laptop so I can reproduce all experiments from my research paper “Evaluating Cross-Dataset Robustness of Machine Learning Models for Malware Detection”.
The setup must support LightGBM, CNN, and GNN experiments exactly as implemented in my project files.

The freelancer will be responsible for installing and configuring:

1. Required Environment

Python 3.10+

Jupyter Notebook

Pip & virtual environment

GPU support (CUDA & cuDNN) if my laptop supports it

Required Python packages including:

pandas, numpy

scikit-learn

lightgbm

seaborn, matplotlib

tensorflow / keras (for CNN)

torch + torch_geometric (for GNN)

scipy

pillow

tqdm

networkx

sklearn.calibration utilities

2. Dataset Setup

Please prepare the following datasets on my laptop:

• EMBER2018 (CSV features)

Used for LightGBM.

• Malimg Dataset

Folder structure with malware family images for CNN classification.

• LAMDA Dataset

CSV with features and labels for GNN and drift analysis.

I will provide the datasets if needed.

3. Code Setup

You will reproduce the exact execution environment from my notebooks (included in the research document):

LightGBM implementation (Feature removal, training, calibration curve, ROC-AUC)

CNN implementation (image preprocessing, augmentation, training, evaluation, calibration)

GNN (GraphSAGE) (graph creation with cosine similarity, training, ROC, MCC, calibration)

You must ensure the code runs without errors.

4. Testing & Verification

After installation, you must:

Run each model and ensure results match expected outputs in my research

Paper_need to be MDPI

Show me the ROC curve, calibration curve, and confusion matrix for each model

Ensure the environment is stable and reusable

5. Deliverables

Fully working environment on my laptop

All dependencies installed

All datasets correctly organized

Jupyter Notebook running all 3 models successfully

A short guide explaining how I can run each experiment again

Ideal Freelancer

Strong experience with Python ML environments

Experience with LightGBM, CNNs, GNNs

Familiar with malware detection datasets

Experience with PyTorch Geometric is a bonus

Able to troubleshoot CUDA/GPU compatibility issues

Patient and able to walk me through the setup step-by-step (remote desktop)
Project Duration

1–2 days setup