Deep Learning Expert for Medical Imaging Code Refinement and Publication
Budget: ₹12,500 – ₹37,500 INR
Title: Experienced Deep Learning developer needed to refine medical imaging Code for Journal Publication
Job Description:
Seeking a highly qualified deep learning researcher/developer to critically review, enhance, and optimize the code for my project to meet the standards of a peer-reviewed journal publication. My work focuses on skin cancer classification using deep learning with the HAM10000 dataset, comparing three architectures: SENet, DenseNet201, and Xception. While the core implementation is complete, I need an expert to address potential shortcomings and improve reproducibility to strengthen the manuscript.
Project Scope & Key Responsibilities:
1. Code Review & Best Practices:
Audit existing code (Python/TensorFlow/Keras) for technical flaws, efficiency, and adherence to reproducibility standards.
2. Methodology Validation:
Verify statistical significance of results.
Check evaluation metrics (e.g., AUC-ROC, sensitivity, specificity, F1-score) and ensure comparisons are fair (same training conditions, hardware, preprocessing).
3. Model Improvements:
Suggest architectural tweaks (e.g., attention mechanisms, hybrid models), hyperparameter tuning strategies, or regularization techniques to boost performance.
Explore explainability methods (e.g., Grad-CAM, SHAP) to strengthen clinical relevance.
4. Reproducibility & Documentation:
Ensure code is fully reproducible with clear READMEs, environment files (Docker/conda), and preprocessing steps.
Assist in creating figures/tables for the manuscript (e.g., confusion matrices, ROC curves, feature visualizations).
5. Ideal Candidate Qualifications:
Hands-on experience with HAM10000 or similar dermatology datasets (e.g., ISIC Archive).
Strong familiarity with SENet, DenseNet, Xception, and their implementations in TensorFlow.
Experience preparing code for academic reproducibility (e.g. GitHub templates).
Ability to communicate technical feedback clearly and collaborate on manuscript revisions.
6. Deliverables:
A detailed report of code improvements, methodological gaps, and performance enhancement strategies.
Refactored, well-documented code with reproducibility checks.
Assistance in responding to reviewer comments post-submission.
Job Description:
Seeking a highly qualified deep learning researcher/developer to critically review, enhance, and optimize the code for my project to meet the standards of a peer-reviewed journal publication. My work focuses on skin cancer classification using deep learning with the HAM10000 dataset, comparing three architectures: SENet, DenseNet201, and Xception. While the core implementation is complete, I need an expert to address potential shortcomings and improve reproducibility to strengthen the manuscript.
Project Scope & Key Responsibilities:
1. Code Review & Best Practices:
Audit existing code (Python/TensorFlow/Keras) for technical flaws, efficiency, and adherence to reproducibility standards.
2. Methodology Validation:
Verify statistical significance of results.
Check evaluation metrics (e.g., AUC-ROC, sensitivity, specificity, F1-score) and ensure comparisons are fair (same training conditions, hardware, preprocessing).
3. Model Improvements:
Suggest architectural tweaks (e.g., attention mechanisms, hybrid models), hyperparameter tuning strategies, or regularization techniques to boost performance.
Explore explainability methods (e.g., Grad-CAM, SHAP) to strengthen clinical relevance.
4. Reproducibility & Documentation:
Ensure code is fully reproducible with clear READMEs, environment files (Docker/conda), and preprocessing steps.
Assist in creating figures/tables for the manuscript (e.g., confusion matrices, ROC curves, feature visualizations).
5. Ideal Candidate Qualifications:
Hands-on experience with HAM10000 or similar dermatology datasets (e.g., ISIC Archive).
Strong familiarity with SENet, DenseNet, Xception, and their implementations in TensorFlow.
Experience preparing code for academic reproducibility (e.g. GitHub templates).
Ability to communicate technical feedback clearly and collaborate on manuscript revisions.
6. Deliverables:
A detailed report of code improvements, methodological gaps, and performance enhancement strategies.
Refactored, well-documented code with reproducibility checks.
Assistance in responding to reviewer comments post-submission.
Related categories:
Python
Research
Artificial Intelligence
Deep Neural Network
Convolutional Neural Network