Welcome! This repository presents my hands-on projects in deep learning and AI, built through continuous self-study and engineering practice. Transitioning from a background in finance and auditing, I’ve focused on practical implementation and deployment-readiness across core AI domains.
This repository showcases a curated portfolio of deep learning projects across four key task types:
- 🖼️ Computer Vision (CV)
- 📚 Natural Language Processing (NLP)
- 📊 Tabular Modeling (Regression, Recommendation)
- 🎨 Generative Models (Diffusion)
Each subfolder includes multiple real-world projects, structured to demonstrate modeling techniques, evaluation strategies, and engineering workflows. This portfolio reflects my journey from finance to AI through hands-on learning and deployment-oriented problem solving.
| Task Type | Folder | Description |
|---|---|---|
| 🖼️ Computer Vision | CV/ |
Image classification, Multi-label learning, Image regression, Optimization techniques |
| 📚 Natural Language Processing | NLP/ |
Named entity recognition, Semantic similarity scoring, Language modeling, Transfer learning, Domain-adaptive process |
| 📊 Tabular Modeling | TabularModeling/ |
Binary-classification, Regression, Tree-based modeling, Feature engineering, Recommendation system, Ensemble strategies |
| 🎨 Generative Models | StableDiffusion/ |
Diffusion models, Image generation, Training strategies |
- Modeling: CNN, RNN, Transformers, Collaborative Filtering, Boosting
- Frameworks: PyTorch, fastai, timm, Hugging Face, scikit-learn, Gradio
- Engineering: Model selection based on data characteristics (e.g., tree vs. NN) , Custom model classes, Low-level gradient descent, Deployment with HF Spaces
- ML Pipeline: Data preprocessing, Augmentation, Feature engineering, Ensembling
- Evaluation: Accuracy, mAP, RMSE, MAE, cross-validation, time-aware splits
- Deployment: Gradio + Hugging Face Spaces for interactive demos
After 8+ years in accounting, audit, and finance, I transitioned into AI by learning from online resources and building a portfolio of applied machine learning projects. My goal is to become a professional deep learning engineer capable of end-to-end solution development.
“From audit reports to model reports, from spreadsheets to tensors.”
If you're reviewing this as a potential employer or collaborator — thank you! I'm always open to connect, discuss ideas, or explore opportunities.