Projects That Drive
Real Results
Explore our portfolio of enterprise solutions that have transformed businesses and created measurable value for our partners.
LinkedIn Content Automation System
An end-to-end LinkedIn automation system powered by Make.com and AI that handles competitor analysis, content ideation, carousel design, video repurposing, and scheduled publishing - completely hands-free.
DigiPromise
DigiPromise AI is a smart AI-powered customer engagement platform that helps businesses automate support, generate leads, and increase sales with intelligent chatbots and AI calling agents. It provides 24 7 conversational automation across websites, social platforms, and ecommerce systems to handle customer inquiries, order tracking, appointment scheduling, product recommendations, and more. The platform can be fully white-labeled so companies can offer advanced AI services under their own brand while improving customer experience and operational efficiency.
Luke - AI Legal Assistant for Peruvian Law
An AI powered legal assistant platform designed to support lawyers, law firms, and legal researchers working with Peruvian law. The system provides an intelligent chatbot that answers legal queries with proper citations, retrieves relevant legal documents, and summarizes laws and case references. It leverages advanced Retrieval Augmented Generation to ensure responses are accurate, traceable, and grounded in verified legal sources, helping users save time while maintaining legal reliability.
NexusFlow
An AI powered agency platform that automates the complete lead to conversion lifecycle. It combines CRM, multi channel lead capture, email and SMS outreach, intelligent lead qualification, booking, analytics, and workflow automation into a single multi tenant system. The platform also supports AI driven calls, proposal generation, customer support, and client portals, enabling agencies to scale operations efficiently while maintaining full manual control.
Automated Conference Agenda Updates
This project focuses on the creation of a highly efficient, AI-driven system that automates the process of updating conference agendas. By leveraging Make.com, CustomGPT.ai, and Google Drive, the system intelligently processes real-time research call data and updates agendas, enhancing the decision-making process and minimizing manual intervention. This solution integrates multiple platforms seamlessly, ensuring optimal efficiency and accuracy during the pre-event phases.
PriceHunt – AI Powered Price Comparison with MCP
Built PriceHunt, a Model Context Protocol (MCP) implementation that helps users find the lowest-priced, high-rated products across major Pakistani e-commerce platforms like Daraz, Telemart, and iShopping. Key Features: - Multi-platform product search - Filters products with 4+ star ratings - Finds the cheapest genuine deals - AI-powered product matching with Google Gemini - Streamlit chat-based interface with memory - Async scraping for faster performance
CodePilot – Coding Assistant for VS Code
Built an AI-powered VS Code extension that transforms the coding experience by integrating AI-powered assistance directly into vs code, to assist developers with tasks like: - Explaining code with full context - Fixing complex bugs in selected code - Generating boilerplate files and code - Chatting with AI using complete file and project context
Mobile MAS – Multi-Agent System for Mobile Phones Analysis
Built an intelligent Multi-Agent System (MAS) that assists users in buying or selling used mobile phones by intelligently coordinating multiple specialized AI agents, combining web scraping, image understanding, and machine learning - all orchestrated through a ReAct-style Supervisor Agent built using LangGraph and Gemini What It Does: - Users can provide a mobile model or even an image of a device, and the system autonomously: - Tells the exact brand and model. - Shows full specs. - Predicts a fair price for selling or buying. - Helps decide if the deal is good or not.
YouTube Videos Agent | LangGraph + Google Veo 3
Developed an automation system that completely automates the YouTube content creation pipeline - from trend analysis to video publication. The system intelligently identifies trending topics, generates creative content, produces videos, and handles distribution with minimal human intervention. Key Features - LangGraph Workflow: Manages full lifecycle - from topic discovery to video publishing. - Interruptible Logic: Pauses for manual approvals(human-in-the-loop) for topics and videos. - Trending Topics: Uses YouTube Data API v3 to fetch hot topics. - Video Generation: Creates 8-second shorts using Google’s Veo 3 (Python SDK). - Streamlit UI: Provides interface for human approvals. - YouTube Automation: Uploads videos with titles/descriptions and logs results. - CSV Logging: Automatically logs video title, description, URL, and upload date/time for tracking.
PrintsStudio – A Full-Stack Blazor Web Application for Custom Printing
Developed a modern, role-based web platform for customizing, designing, and ordering printed materials like wedding cards, business cards, and personalized gifts. The system offers real-time previews, dynamic product customization, and a robust admin dashboard - all built using Blazor WebAssembly and .NET Core.
License Plate Reader
License Plate Detector & Reader using YOLOv5 + EasyOCR Built a computer vision system that detects and reads vehicle license plates from both images and videos - ideal for real-world use cases like traffic surveillance, smart parking, and automated toll collection. Key Highlights: - YOLOv5 for fast and accurate license plate detection - EasyOCR for extracting alphanumeric text from plates - Supports both images and real-time videos - Trained on a real-world annotated dataset from Roboflow Universe - Automatically saves outputs with bounding boxes and recognized text overlays ~The pipeline detects license plates using a custom-trained YOLOv5 model, crops the detected region, and runs OCR with EasyOCR to retrieve the plate number — all handled seamlessly in one script.
CodeQA — RAG-based AI Chat with Your GitHub Repositories
CodeQA is an AI-powered developer assistant that allows users to chat with any public GitHub repository and get contextual answers about its structure, logic, and implementation using Retrieval-Augmented Generation (RAG). I built this project as part of my learning journey into RAG pipelines and LLM applications, aiming to gain hands-on experience in building intelligent developer tools. The app fetches a GitHub repo, breaks down code intelligently, embeds it using semantic vector representations, and answers user queries using a powerful LLM (DeepSeek/GPT). Key Features: - Load any public GitHub repo by just pasting the URL - Automatic code chunking using language-aware splitting - Semantic embeddings with HuggingFace models (open-source) - Vector search via ChromaDB - RAG-powered chat with conversation memory