Aegis is a progressive web app built to manage monthly security contributions in small residential vicinities. It replaces WhatsApp based tracking and Excel audits with a clear household level billing cycle, verified payment status, and push notification reminders. The project focuses on reducing manual reconciliation, removing social friction, and iterating based on real world usage rather than assumptions.

SubZup is a subscription manager I am building to solve a problem most of us quietly deal with. Too many subscriptions, scattered across emails, cards, and apps, and no simple way to see them all at once. SubZup brings everything into one calm, beautifully designed place so you instantly know what you are paying for. The idea is to smartly read your email inbox to detect existing subscriptions and use AI to help you add and manage new ones.

RAG with MongoDB & LlamaIndex A powerful implementation of Retrieval-Augmented Generation that turns a MongoDB collection into a searchable knowledge base. By integrating vector search with LlamaIndex and OpenAI, this project enables real-time, context-aware AI interactions with private datasets.

BFF Optimization Lab: NestJS & Request Coalescing > Solved the problem of redundant upstream API calls in microservice architectures by building a BFF that "coalesces" overlapping requests. This architecture ensures that even during traffic spikes, identical requests share a single backend fan-out, significantly reducing server load and latency. Includes a built-in "bombard" script to prove performance gains through real-time metrics and custom headers (X-Cache, X-Coalesce).

This is a shared expense app built to simplify how people manage money together. It helps friends couples and families track who paid what split expenses fairly and always see clear balances. Instead of forcing instant settlements it keeps a running overview so money can be settled naturally. The focus is on clarity simplicity and removing awkward money conversations from everyday life.