🎯 About Me 🎓 3rd-year BS Information Technology student at University of Sindh, Jamshoro (6th semester) 🤖 Specializing in LLM-integrated platforms, agentic AI systems, and cloud-deployed backends 📜 Hold 8 Anthropic-verified credentials, including Model Context Protocol (MCP) — Apr 2026 🏛️ Completed two government-backed internships at PITP–MUET (Government of Sindh) in Cloud Computing and Graphic Design 🔬 Currently deepening expertise in agentic workflows (LangGraph) and n8n automation 🌍 Based in Hyderabad, Pakistan

🚀 What I Do 🤖 Agentic AI Systems — Building tool-calling assistants with LangGraph, Mistral AI, and persistent vector memory (pgvector) ⚙️ Backend Engineering — Designing REST APIs and relational schemas with Flask, PostgreSQL, and Supabase 🔐 Machine Learning for Security — Training XGBoost/Random Forest models for network intrusion detection with SHAP explainability ☁️ Cloud Deployment — Shipping full-stack platforms on Railway and Vercel with CI/CD pipelines

AI & LLM Engineering

Machine Learning

Backend & Cloud

Frontend

Currently Exploring

A fully agentic AI assistant, not just a chatbot wrapper

Built a LangGraph-based agentic assistant with real tool-calling — email, calendar, and reminders — backed by persistent vector memory (pgvector) for contextual recall across sessions. Implemented a real-time streaming chat interface (Flask-SocketIO) with a live agent-reasoning trace panel exposing the model's tool-selection steps, plus voice I/O via Piper TTS and a custom clap-to-wake system. Deployed a public sandbox (Railway + Vercel) to demo the architecture without exposing the private integrations.

Social prediction platform for FIFA World Cup 2026

Built a full-stack social platform with a 6-module REST API and 5-table relational schema, where users post match predictions and an automated AI-scoring pipeline (APScheduler + Mistral AI) grades them against live results and generates real-time roasts — eliminating manual scoring entirely. Integrated live football data via the API-Football feed.

~99% accuracy ML pipeline on real-world cybersecurity data

Built an end-to-end ML pipeline on the CICIDS-2017 dataset, classifying 14 network attack categories (DDoS, Port Scan, Botnet, Brute Force) with XGBoost achieving ~99% accuracy and a near-zero false-positive rate. Applied SHAP for model explainability, identifying the top predictive features driving each classification, and packaged the pipeline as a reproducible module for one-command reuse.

Production-grade personal platform on multi-cloud free-tier infrastructure

Delivered a Flask + Neon PostgreSQL + Railway platform with 4 REST API endpoints, GitHub CI/CD auto-deploy, dark/light mode, and admin-authenticated contact submissions — built as a capstone project during a government-backed Cloud Computing internship.

Cloud Computing Intern — PITP–MUET, Jamshoro (Govt of Sindh) · Mar – May 2026 Delivered a capstone Flask + PostgreSQL + Railway platform with CI/CD as part of a government-backed technical training program, and reduced lab environment setup time by ~30% by scripting VM configuration for trainee onboarding.

Certified Graphic Designer Intern — PITP–MUET (Govt of Sindh) · Jan – Mar 2025 Produced 10+ branded visual assets for institutional print and digital communications under a government-accredited design program.

Relevant coursework: AI & Machine Learning, Advanced DBMS, Cybersecurity, Cloud Computing, Computer Networks

🔭 Current Mission 🔭 Building: Alfred's core agent architecture (LangGraph + vector memory) into a production-ready private build 🌱 Exploring: Agentic workflow design, n8n automation, and multi-hop LLM tool orchestration 🤝 Open to: AI/ML internships, backend engineering roles, and collaborations on agentic AI projects 💬 Ask me about: LangGraph, Mistral AI, Flask backends, or intrusion detection with XGBoost ⚡ Fun fact: I built a clap-to-wake assistant using real-time FFT spectral analysis, no cloud needed 🎯 Goal: Shipping AI systems that are genuinely useful, not just impressive demos

Open to internships, collaborations, and interesting problems.

⭐ Star a repo if my work impressed you ⭐