Muhammad
Hassan Irshad.
AI/ML Engineer focused on building intelligent, production-oriented applications with Large Language Models, AI agents, automation and Python.
Selected Works
AI agents, LLM applications, automation, and data-driven engineering for real-world problems.
Cyber-Threat Hunter
An autonomous AI-powered security system that uses a Red/Blue/Green agent architecture to detect, validate, and automatically remediate application vulnerabilities.
Smart Location Optimization
AI-driven location intelligence and mathematical optimization for business placement.
AI-Powered WhatsApp Message Routing
LLM-based message classification, routing, and workflow automation.
Technical Stack
Tools and technologies used across my projects.
AI / MACHINE LEARNING
AGENTIC AI
PROGRAMMING
BACKEND / DATA
OPTIMIZATION / GEOSPATIAL
TOOLS / OBSERVABILITY
AI ENGINEERING FOCUS
Building applications around large language models, structured prompting, API integration, and intelligent workflows.
Designing multi-agent workflows where specialized agents collaborate to analyze problems, make decisions, and execute actions.
Exploring retrieval-based AI workflows that ground model outputs in external data and tools.
Building systems with validation, observability, error handling, structured outputs, and automated workflows.
Project Screenshots
6 views
Cyber-Threat Hunter & Auto-Remediator
An autonomous AI-powered security system using a Red/Blue/Green agent architecture to detect, validate, patch, and re-verify application vulnerabilities.
The Problem
Traditional security tooling can identify potential vulnerabilities but may leave false positives and manual remediation work for developers.
The Solution
A closed-loop multi-agent workflow combines automated scanning, runtime validation, intelligent patch generation, syntax validation, Git commits, and re-verification.
Key Capabilities
- Red Agent: vulnerability discovery, secret detection, AST analysis, dependency scanning, STRIDE/MITRE classification.
- Blue Agent: runtime validation with simulated requests and OpenTelemetry traces.
- Green Agent: similarity cache, deterministic templates, Gemini LLM fallback, AST validation, Git-based patching.
- Closed-loop re-scan with a two-attempt remediation cap and human escalation.
About
AI/ML engineering with a focus on useful, automated systems.
I’m Muhammad Hassan Irshad, a Computer Science graduate focused on AI/ML engineering and intelligent software systems.
My current focus is building practical AI applications using Python, LLMs, agentic workflows, and automation. I enjoy turning complex problems into systems that can reason, analyze data, make decisions, and execute actions.
My recent work includes a multi-agent Cyber-Threat Hunter and Auto-Remediator, a Smart Location Optimization System, and AI-powered message routing. I’m particularly interested in LLM applications, RAG, agent orchestration, and production-oriented AI systems.
Education
NFC Institute of Engineering, Faisalabad
BS Computer Science · 2022–2026
Minor: Mathematics
Experience
Professional experience and applied engineering work.
Contributed to DUKE Energy and real-world software development projects.
Participated in project development, technical discussions, meetings, and brainstorming sessions.
Collaborated with supervisors and team members to complete assigned development tasks.
Gained practical experience with software development workflows, organizational processes, and IT solutions.
Developing AI-powered applications involving LLMs, intelligent agents, automation, cybersecurity analysis, and optimization systems.