Muhammad Hammad Irfan
|
Bridging robust backend architecture with secure AI-driven logic. Specialized in orchestrating autonomous workflows, enforcing LLM security, and building scalable, production-ready systems.
About Me
I’m an AI Engineer and Full-Stack Developer specialized in building production-grade autonomous systems. My core focus is bridging scalable backend architecture with secure AI-driven logic, particularly through multi-agent orchestration and enforcing strict LLM tool boundaries. Whether I am auditing enterprise agents, developing vector-based data platforms, or engineering reliable automation pipelines, I build robust systems that solve complex, real-world business constraints.
Multi-Agent AI Orchestration
Building autonomous agent workflows, implementing reliable function calling, and engineering custom AI integrations.
LLM Security & Architecture
Auditing agents against the OWASP Top 10, enforcing Zero Trust tool boundaries, and mitigating Excessive Agency.
Backend & Vector Databases
Designing scalable data flows and semantic search architectures using PostgreSQL, Supabase, and PGVector.
Production Pipelines
Engineering end-to-end automation workflows that bridge complex APIs to solve direct operational business bottlenecks.
Skills & Expertise
Core Stack
AI Engineering & Security
Research & ML Systems
Backend & Databases
Cloud & Infrastructure
Architecture & Engineering
Featured Research & Engineering Projects
A selection of recent peer-level research, AI security engines, clinical decision support systems, and production platforms.
AI-Powered Applications
A multi-agent AI system that connects to GitHub to autonomously map dependencies, detect vulnerabilities, and optimize infrastructure spend.
.png&w=1920&q=85)
Research Projects
An explainable clinical decision-support system for nephrology, built with Kyung Hee University Hospital, classifying patients into 33 physician-validated clinical situations.
AI-Powered Applications
An AI-powered hiring platform that maps resumes to job descriptions using vector databases and generates automated skill gap analyses.
Web & PWA Products
Offline-first multi-business management and invoicing PWA with seamless background cloud sync, multi-tenant switching, and PDF billing.
Production Full-Stack Systems
A production-scale university automation platform built to streamline communication, scheduling, and academic workflows.
AI-Powered Applications
An AI-powered English speaking companion that helps users practice real conversations through voice, not just text or grammar drills.
Research Projects
An open-source adaptive red-teaming engine for agentic AI and RAG systems that implements published attack research and outperforms fixed-checklist approaches by re-ranking attacks against a learned evidence model.
Research Projects
Contributing to a £2.8M UKRI-funded consortium building a digital-twin model of a refrigerant-free evaporative cooling system, combining first-principles equations with ML surrogate models.
Experience
Forward-Deployed AI Engineering Intern
DevNeuron — Sunnyvale, California, US
- Building Aginiti-Redteam, an open-source adaptive red-teaming engine for agentic AI and RAG systems, implementing four attacks from peer-reviewed security research (ICLR, ACM CCS, IEEE TIFS) covering data reconstruction, membership inference, and system-prompt extraction.
- Designed a benchmark target with multiple independently-toggleable defense layers and validated the engine head-to-head against random and fixed-order baselines and NVIDIA's garak scanner, an industry-standard tool.
- Built a three-stage verification pipeline so no security finding is accepted on a single model's opinion, combining deterministic checks, LLM-based judgment, and independent confirmation against real source data.
AI & Software Developer
KidneyCare Clinical Assistant — SEECS, NUST
- Developed a clinical decision support system to assist nephrologists in analyzing patient data and treatment options.
- Implemented rule-based AI logic aligned with medical guidelines to generate safe, explainable recommendations.
- Designed workflows that handle missing or unreliable lab data, ensuring reliable and structured outputs for physicians.
- Validated the system against historical patient records from Kyung Hee University Hospital, confirming its recommendations were consistent with actual nephrologist decisions.
Research Contributor (Final Year Project)
S2Cool — UKRI Ayrton Challenge Consortium, NUST SEECS
- Contributing to a digital twin of a refrigerant-free evaporative cooling system as part of a £2.8M international UKRI-funded research consortium led by Northumbria University.
- Building machine learning surrogate models trained on real sensor data from physical prototype testing, combined with first-principles heat and mass-transfer modeling.
Web Developer Intern
AskUsSolution (Remote)
- Contributed to small-scale web projects, gaining practical experience with React, Node.js, REST APIs, and Git workflows.
- Developed frontend and backend features, learned deployment basics, and strengthened understanding of production-ready web development.
Education
National University of Sciences and Technology (NUST)
Bachelor of Science in Computer Science
Islamabad, Pakistan
Aspire College
Intermediate in Pre-Engineering
Get In Touch
Have a project in mind, research collaboration, or inquiry? Send me a message and it will be delivered directly to my inbox at muhammadhammadirfan1@gmail.com.
Direct Email
muhammadhammadirfan1@gmail.comLocation
Islamabad, Pakistan
Availability
I'm currently open to research collaborations, AI engineering roles, and innovative freelance projects.