Co-founder and AI engineer at KortexMail.

Muhammad Bilal  • Peerlist

KortexMail is an agentic email platform built for professionals who are overwhelmed by inbox noise. It understands the full context of your work, including threads, relationships, deadlines, and priorities, to autonomously read, organize, and act on your emails, surfacing only what truly needs your attention.Unlike traditional email clients that simply display messages, KortexMail functions as an intelligent assistant that drafts replies in your tone, extracts tasks and events, ranks emails by importance, and connects with tools like Google Calendar and Notion. Built on the principles of agentic intelligence, contextual clarity, and reclaimed time, it gives back the hours your inbox has been stealing. Join those who have chosen focus over noise and clarity over clutter. The quietest inbox you'll ever own.

Kortex Mail — Agentic, Context-Aware Email

**Heart Disease Prediction: A Comprehensive ML Pipeline** I'm excited to share my latest project - a robust machine learning system for heart disease prediction that implements industry best practices and compares multiple algorithms to find the optimal solution. **Technical Highlights:** • **Multi-Model Comparison**: Evaluated 5 classification algorithms (Random Forest, Gradient Boosting, Logistic Regression, SVM, KNN) • **Advanced Pipeline Architecture**: Proper preprocessing with separate handling for numerical, ordinal, and categorical features • **Data Leakage Prevention**: Correct train-test splitting and pipeline implementation • **Class Balancing**: SMOTE technique for handling imbalanced datasets • **Robust Validation**: 5-fold stratified cross-validation for reliable performance metrics **Key Features:** • Comprehensive data preprocessing (imputation, scaling, encoding) • Multiple evaluation metrics (Accuracy, Precision, Recall, F1-Score) • Feature importance analysis • Automated model selection based on performance • Detailed classification reports and confusion matrices **Results:** The pipeline automatically selects the best-performing model through rigorous cross-validation, achieving strong predictive performance on heart disease classification. **Best Practices Implemented:** - Scikit-learn pipelines for reproducibility - Stratified sampling for imbalanced data - Proper evaluation methodology - Modular, extensible code structure This project demonstrates my commitment to building production-ready ML solutions that follow industry standards. The complete code and documentation are available on my GitHub. What's your experience with healthcare ML applications? I'd love to connect with others working in this space! #MachineLearning #Healthcare #DataScience #HeartDisease #Python #ScikitLearn #PredictiveAnalytics #MLOps #HealthTech

Heart_disease_prediction

House_price_prediction

# Wikipedia Search Agent - LinkedIn Project Description **Project Title:** AI-Powered Wikipedia Search Agent **Description:** Developed an intelligent conversational AI agent using LangChain and OpenAI's GPT-3.5-turbo that can search Wikipedia and maintain conversation context. The agent implements the ReAct (Reasoning and Acting) pattern to autonomously decide when to search for information and provide accurate, contextual responses. **Key Features:** • Built with LangChain framework for robust agent orchestration • Integrated Wikipedia REST API for real-time information retrieval • Implemented conversation memory to maintain context across interactions • Used ReAct pattern for intelligent decision-making and tool usage • Added comprehensive error handling and logging for production readiness **Technical Stack:** • Python, LangChain, OpenAI API • Wikipedia REST API integration • Environment variable management with python-dotenv • Structured logging and exception handling **Impact:** Created a foundation for intelligent information retrieval systems that can be extended to multiple data sources. The modular architecture allows easy integration of additional tools and APIs, making it suitable for various AI-powered applications. **Skills Demonstrated:** • AI Agent Development • API Integration • Natural Language Processing • Python Programming • Error Handling & Logging • ReAct Pattern Implementation **GitHub:**https://github.com/bilalsjutt/Simple_wikipedia_search_Agent