Chat
Hello, I’m Hudson Mathew, a versatile developer specializing in Python, frontend and backend engineering, and UI/UX design. I enjoy turning ideas into reliable, efficient, and visually polished products that solve real problems and deliver smooth user experiences.
Read MorePython development services focused on building clean, scalable, and maintainable applications. This includes creating automation scripts, developing Django-based backend systems, and handling data processing workflows. REST APIs are built and integrated to connect apps, databases, and third-party services. AI/API integrations are also implemented, including chatbot-style features that add smart functionality to products.
Frontend development services focused on building clean, responsive, and user-friendly web interfaces. I create well-structured pages with HTML, style modern layouts with CSS, and add interactivity using JavaScript. I build reusable UI components like navbars, cards, forms, and buttons with consistent design across pages. I also integrate frontend with backend systems (like Django templates and APIs) to deliver dynamic, real-world web apps.
Backend development services focused on building secure, scalable, and high-performance systems. I develop application logic, create REST APIs, and connect applications to databases for reliable data handling. I implement authentication, authorization, validation, and security best practices, and integrate third-party services to support real-world features and smooth deployments.s.
UI/UX design services focused on creating clean, modern, and easy-to-use digital experiences. I design layouts, components, and visual systems that stay consistent across screens and features. I improve user flows, navigation, and accessibility so products feel simple and intuitive to use. I focus on responsive, user-centered design that enhances usability and overall product quality.
Hero’s AI is a Django-based, multi-modal personal assistant that orchestrates multiple LLM providers (Google Gemini + OpenRouter, with automatic fallback) to deliver reliable text chat, coding help, voice responses, web search, and file Q&A. A custom NLP pipeline cleans input, detects intent with regex-driven routing, and selects the best handler/model per task while adapting token budgets for speed vs depth. Its Infinsight module adds RAG-powered analytics for CSV/Excel/PDF: documents are chunked, embedded (gemini-embedding-001), stored in Pinecone, and retrieved via semantic search to ground answers in user data. For quantitative questions, the system generates and safely executes Python/Pandas to produce analyst-style reports (trends, forecasting, correlations). With per-user encrypted API key storage, Google OAuth, and persistent chat sessions, Hero’s AI showcases end-to-end LLM, NLP, RAG, and ML engineering in a production-ready web app.
Infinsight is the RAG-powered data analytics layer inside Hero’s AI, built to turn uploaded files into queryable knowledge. Users can upload CSV, Excel, or PDF reports; the system parses and chunks content (row summaries for tables, page-level chunks for PDFs), then generates embeddings with gemini-embedding-001 and stores them in Pinecone under a user-specific namespace. On each question, Infinsight performs semantic retrieval to surface the most relevant chunks, injects that grounded context (plus dataset schema) into the LLM, and produces answers that stay tied to the user’s data. For quantitative requests (totals, trends, forecasting, correlations), it generates Python/Pandas code and executes it safely in a sandboxed interpreter, then converts results into a clear, analyst-style report. The result is accurate, explainable, natural-language analytics without manual coding.
EpicOutlet (Django Project) EpicOutlet is a full-stack e-commerce web application built with Python and Django that lets users browse, search, and purchase products across categories like electronics, fashion, groceries, and home items. It includes user authentication, a shopping cart, a favourites list, and an AI shopping chatbot (Shop-Bot) that helps users find products through natural-language queries (for example, “phones under ₹30,000”) by searching the product database with filters like category and price range. The system follows Django’s MVT architecture, stores data in PostgreSQL (Supabase), hosts images on Cloudinary, serves static files with WhiteNoise, and is deployed on Render using Gunicorn. It also provides a REST API (via django-tastypie) under /api/v1/ for external access to product and related data. Overall, the project demonstrates end-to-end web development plus modern AI integration and cloud deployment in a production-ready setup.
Spaceship Titanic (Weighted Ensemble Model) is a Kaggle machine learning project that predicts whether passengers were “Transported” after a spacetime anomaly collision. It uses features like age, home planet, destination, cabin details, and amenity spending to clean missing values, engineer cabin-based features (Deck/Number/Side), encode categories, and scale data for training. Multiple models (Logistic Regression, Random Forest, XGBoost, CatBoost, AdaBoost, and a Deep Neural Network) are trained and combined using a weighted ensemble to improve accuracy over any single approach. The final system produces a submission-ready CSV and is evaluated using standard classification metrics, showing strong preprocessing, balanced model contribution through weighting, and improved prediction reliability on unseen test data.