A team of CS graduates from BZU Multan & Arid University Rawalpindi โ engineering production-ready AI systems, cloud infrastructure, and intelligent automation that solve real problems.
Our team clinched 1st Place at the Innovista Agentic AI Hackathon 2025 by building RescueLink โ an AI-powered emergency response and automation platform. The project impressed judges with its real-world applicability, combining NLP-driven triage, geolocation prioritization, and multi-step agentic workflows to coordinate emergency response at scale.
CS graduates from Bahauddin Zakariya University Multan & Arid University Rawalpindi, building production-grade AI systems, cloud platforms, and automation solutions that go beyond academia into the real world.
Each of us brings a distinct specialization โ together we cover the full stack from AI infrastructure to agentic systems, automation, and intelligent applications.
From hackathon winners to live SaaS platforms โ everything we build runs in production and solves real problems.
AI-powered emergency response & automation platform that won 1st Place at Innovista Agentic AI Hackathon 2025. Features event-driven triage workflows, NLP-based incident classification, geolocation prioritization, and automated resource dispatch โ coordinating emergency response at scale with agentic AI.
End-to-end autonomous vehicle control using deep learning. A CNN inspired by NVIDIA's self-driving architecture predicts steering angles directly from front-facing camera images โ trained on the Udacity Behavioral Cloning dataset. No manually programmed rules; the model learns lane markings, road curvature, and driving direction autonomously.
Live student SaaS at sageai.live that converts PDF/PPTX notes into AI-generated audio lectures. Running on AWS with a free tier (5 lectures/day) serving real users โ architected to run under $1/day through deliberate cost optimization using Groq API, Edge TTS, and SQLite over RDS.
End-to-end automated content pipeline: scrapes Reddit stories, generates AI scripts and voiceovers via Gemini API + Edge TTS, and produces synchronized subtitle videos with AMD AMF hardware encoding. Decoupled architecture with FastAPI backend communicating via Amazon SQS for reliable concurrent video generation.
Real-time American Sign Language recognition achieving 92%+ accuracy using computer vision and deep learning. Built with MediaPipe hand landmark detection and a custom TensorFlow model trained on the Microsoft gesture dataset. Enables accessibility for the hearing-impaired community.
Deployed 15+ business process automations using Power Automate and Make.com for real clients โ invoice processing, lead management, approval workflows, and report generation. Achieved 60% reduction in manual processing time and 98% automation reliability with robust error-handling and retry logic.
We work with startups and organizations to build intelligent, scalable solutions across AI, cloud, and automation.
Design and deployment of LLM-powered autonomous agents with tool calling, RAG pipelines, and multi-step decision workflows for complex task automation.
Production-grade AWS deployments with cost optimization, high availability, distributed task scheduling, and SRE monitoring practices.
End-to-end workflow automation using Power Automate, Make.com, and custom APIs โ reducing manual effort and improving operational reliability.
Object detection (YOLOv8/v9), gesture recognition, real-time classification systems, and custom ML model training for domain-specific applications.
Cross-platform mobile apps in Flutter/React Native and web applications with React/TypeScript โ from concept to production deployment.
Security-first development incorporating OWASP guidelines, vulnerability assessment, secure API design, and role-based access control systems.
Whether you have a project idea, collaboration opportunity, or just want to say hello โ we'd love to hear from you.
We're a team of four based in Multan & Rawalpindi, Pakistan โ open to freelance projects, startup collaborations, and full-time opportunities globally.