Building scalable AI solutions and state-of-the-art machine learning models to solve complex real-world problems.
- 🎓 BS Data Science at GIFT University (Specializing in Computer Vision & ML).
- 🧠 Focused on Deep Learning, Domain Adaptation, and Foundation Models.
- 🚀 Built and fine-tuned architectures like ConvNeXt, Prithvi-EO, and Llama 3.
- 💡 Proven track record of improving SOTA benchmarks (e.g., +23.3% F1-score on deforestation detection).
- 🌐 Currently building scalable RAG architectures and working as a Freelance AI Developer.
| Research Project | Focus | Impact & Links |
|---|---|---|
| Fourier Domain Adaptation & Evidential DL | Cross-domain deforestation detection replacing adversarial training with parameter-free FDA. | 📈 +23.3% F1-score improvement. 📝 Paper Under Review 💻 Source Code |
| Land Cover Mapping with Foundation Models | LoRA fine-tuning on the 300M NASA/IBM Prithvi-EO model using Sentinel-2 imagery. | 🌍 Superior zero-shot cross-city generalization. 📊 Quantified +21.3% urban expansion. 💻 Source Code |
| Project | Description | Tech Stack |
|---|---|---|
| AskDoc AI | Enterprise-grade RAG Chatbot allowing real-time PDF querying with semantic search. | FastAPI, LangChain, FAISS, Llama 3 |
| Market Basket Analysis | Discovered hidden purchasing patterns on 540k+ transactions using Apriori and FP-Growth. | Python, Pandas, Data Mining |
| Automated Data Pipeline | Scalable web scrapers deployed as REST APIs for unstructured e-commerce data extraction. | Node.js, Playwright, REST APIs |
"Pushing the boundaries of what's possible with Artificial Intelligence."