Skip to content
View Shweta-Portfolio's full-sized avatar
🎯
Focusing
🎯
Focusing
1 follower · 0 following

Block or report Shweta-Portfolio

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Shweta-Portfolio/README.md

Shweta Debjit Sarkar

Independent ML Researcher | Data Scientist | Machine Learning Engineer | Business Analytics Graduate

Typing SVG

About Me

Data Scientist and ML Researcher focused on building interpretable, responsible AI for real-world impact. MSc Business Analytics & International Business (University of Dundee). Four peer-reviewed preprints in clinical and health informatics. Currently building a deep learning portfolio spanning computer vision and NLP.

What I work on:

  • Interpretable ML — SHAP, Grad-CAM, explainability pipelines
  • Computer Vision — transfer learning, wildlife detection, ResNet
  • NLP & Ethical AI — bias detection, text analysis at scale
  • Predictive Analytics — clinical risk stratification, churn modelling, survival analysis
  • Research — 4 DOI-indexed preprints on Zenodo

Core Competencies:

  • Data Analysis & Visualization
  • Machine Learning & Predictive Modeling
  • Business Intelligence & Dashboard Development
  • Process Optimization & Workflow Mapping
  • Cross-functional Collaboration

Technical Skills

python mysql pandas scikit_learn seaborn powerbi jupyter excel

Connect

LinkedIn
Email

Profile views

Pinned Loading

  1. -Coral-Reef-Health-Classification -Coral-Reef-Health-Classification Public

    Binary coral reef health classifier (healthy vs bleached) using VGG16 and EfficientNetB0 transfer learning on 923 images. Demonstrates bias-variance tradeoff VGG16 overfit with 119M params while fi…

    Jupyter Notebook

  2. Ai-Agency-Student-Learning Ai-Agency-Student-Learning Public

    Pilot observation study examining how secondary school students (Years 7–11) engage with AI-generated answers in the classroom and whether students who actively interrogate AI output show higher co…

    Jupyter Notebook

  3. Clinical-RAG-Parkinsons Clinical-RAG-Parkinsons Public

    Clinical RAG pipeline for question-answering over Parkinson's and Alzheimer's research literature

    Jupyter Notebook

  4. Skin-Lesion-Classifier Skin-Lesion-Classifier Public

    7-class skin lesion classifier using ResNet50 transfer learning on HAM10000. Achieves 83.8% balanced accuracy with weighted loss, Grad-CAM interpretability, and per-class ROC curves. Built as part …

    Jupyter Notebook

  5. Skin-Lesion-Classifier-Using-CNN Skin-Lesion-Classifier-Using-CNN Public

    Custom CNN trained from scratch on HAM10000 for 7-class skin lesion classification. Compares baseline CNN V1 (64×64, 3 blocks, 39.5% balanced acc) vs optimised CNN V2 (128×128, 4 blocks, 42.1% bala…

    Jupyter Notebook

  6. Wildlife-Camera-Trap-Detector Wildlife-Camera-Trap-Detector Public

    Wildlife camera trap classifier using transfer learning (ResNet18) + Grad-CAM interpretability. Built with PyTorch. 85%+ test accuracy on Serengeti dataset.

    Jupyter Notebook