AI | ML engineer focused on understanding AI from first principles. I implement algorithms from scratch, fine-tune transformers, and build explainable systems for real-world applications.

What drives me:

• 🧠 Understanding of ML fundamentals through implementation from scratch
• 🛡️ Building explainable AI systems for critical decision-making
• 🚀 Creating production-ready solutions that solve real problems
• 📚 Continuous learning and sharing knowledge with the community

Current Focus:

• Fine-tuning transformer models for cybersecurity threat detection
• Building explainable AI systems with interpretability
• Implementing ML algorithms from scratch to strengthen fundamentals
• Exploring MLOps and scalable deployment strategies

AI-Powered Cybersecurity Threat Detection System

• Fine-tuned DistilBERT and TinyBERT transformers for malware detection achieving 97.6% accuracy
• Implemented ensemble learning (soft voting + stacking) to reduce false positives
• Built interactive Next.js dashboard with LLaMA-3 powered explanations for security decisions
• Tech: PyTorch, Transformers, Ensemble Learning, Next.js, Groq LLaMA-3 API
• Live Demo | Frontend Repo

AI-Powered Bioinformatics Application

• Integrated AlphaFold2 pipeline for 3D protein structure prediction from amino acid sequences
• Built interactive molecular visualization using Py3Dmol
• Deployed full-stack web application with Streamlit
• Tech: Python, Streamlit, BioColabFold (AlphaFold2), Py3Dmol
• Live App

Content-Based NLP Recommender

• Implemented content-based filtering using NLP (BoW, TF-IDF, Cosine Similarity)
• Integrated TMDb API for real-time movie metadata and posters
• Deployed interactive web app with Streamlit
• Tech: Python, NLP, Scikit-learn, Streamlit, TMDb API
• Live Demo

ML Classification for Public Health

• Built classification pipeline for water quality assessment using physicochemical parameters
• Achieved 92.5% accuracy with Random Forest ensemble
• Demonstrated ML applications in automated public health decision-making
• Tech: Python, Scikit-learn, Pandas, Seaborn

Ensemble Regression Models

• Implemented and compared Linear Regression, Decision Trees, Random Forest, and AdaBoost
• Performed comprehensive EDA, feature engineering, and outlier handling
• Random Forest achieved superior performance through ensemble learning
• Tech: Python, Scikit-learn, Pandas, NumPy, Matplotlib

Building Fundamentals

• Implementing machine learning algorithms from the ground up using only NumPy
• Part 1: Linear Regression with mathematical derivations and detailed documentation
• Focus on understanding the mathematical foundations and algorithmic principles
• Tech: Python, NumPy

Machine Learning & Deep Learning

• Neural Networks (from scratch implementation to production deployment)
• Transformer Models (Fine-tuning BERT, DistilBERT, TinyBERT)
• Ensemble Methods (Stacking, Soft Voting, Random Forest, AdaBoost)
• Computer Vision and Natural Language Processing

Explainable AI & Model Interpretability

• Building transparent and interpretable ML systems
• LLM-powered explanations for model decisions
• Feature importance analysis and SHAP values

Software Engineering & Deployment

• Full-stack web development (React, Next.js, Streamlit)
• RESTful API design and integration
• Cloud deployment and scalability (Google Cloud Platform)
• Version control and collaborative development