Fahad Hafeez

Fahad Hafeez

AI Researcher • Founder • OSS Builder

Building trustworthy AI systems, datasets and tools.

Independent AI researcher and founder of Khwand AI, building open-source models, datasets, and research at the intersection of machine learning, cybersecurity, and trustworthy AI.

Trustworthy AI ML Systems Open-source tooling Academic foundation in CS & math
Explore Research View Datasets Download CV

Interests: trustworthy AI, reproducible research, practical ML systems, and building tools that make complex technology more reliable and useful.

Models
Datasets
Publications
Projects
GitHub Stars
Founder, Khwand AI
0 + Publications
0 + Datasets
0 + Projects
0 + GitHub Stars
Founder, Khwand AI

Building Khwand AI

Khwand is an assurance platform for AI agents, helping teams test, harden, and monitor agentic systems before they ship.

Reliability First

Turn brittle agent workflows into production-ready systems with automated checks, regression testing, and safer deployments.

Stability Score

Quantify how ready an agent is to ship with a clear score that surfaces weak spots, failure modes, and reliability gaps.

Continuous Assurance

Move beyond ad-hoc debugging with monitoring, traceability, and resilience checks that keep agents dependable over time.

Built for Shipping Teams

Designed for founders, platform engineers, and AI teams who need practical guardrails without slowing delivery.

Visit Khwand

Open Source Datasets

Datasets published on Hugging Face for experimentation, benchmarking and model development.

cissp-llmbench

Named Entity Recognition Cybersecurity

Cybersecurity-focused benchmark dataset for identifying threats, vulnerabilities and attack vectors in text.

Open Open dataset

safecalib-bench

Safety & Calibration Evaluation

Benchmark dataset for evaluating safety, calibration and robustness of language models in real-world settings.

Open Open dataset

Research

Core areas of investigation and contribution.

  • Trustworthy AI
  • Cybersecurity
  • Large Language Models
  • Agentic AI
  • Federated Learning
  • Explainable AI

Licenses & Certifications

A selection of credentials spanning AI engineering, data science, NLP, cybersecurity, and cloud foundations.

AI Engineer for Developers Associate Data Engineer AI Engineer for Data Scientists Associate Reinforcement Learning in Python Deep Learning in Python Natural Language Processing in Python NLP in Python Anthropic MCP & Claude Foundations AWS Generative AI Google Generative AI HubSpot Email Marketing CompTIA Network+ CISSP Foundational C# with Microsoft

Education & Leadership Background

My academic path combines formal study in artificial intelligence and data science with early leadership and service experience.

University of Lancashire

BSc, Artificial Intelligence and Data Science
2025 – 2028
First-year modules included Computing Challenge, Systems Analysis and Database Design, C++ Programming, Machine Learning Foundations, Ethics, Data and Professionalism, and Mathematics and Statistics for AI and Data Science.

IU International University of Applied Sciences

Bachelor of Applied Science – BASc, Artificial Intelligence
Sep 2024 – Aug 2025

Nelson and Colne College

A Levels in Computer Science, Business, Graphic Design, and EPQ
Sep 2022 – Aug 2024
Grades: A, A, A, D.

Royal Air Force Air Cadets

Served as a Cadet and later Cadet Lance Corporal through the Combined Cadet Force at Maiden Erlegh School in Reading. This experience strengthened my leadership, discipline, reliability, problem-solving, and time-management skills.

Earlier education

Maiden Erlegh School in Reading — GCSEs (Sep 2019 – Aug 2022), grades: 6, 6, 5, 5, 4, 4, 4, 4 M1.
Roots Millennium Schools & Colleges – TME — River Tree Campus, Nowshera (Dec 2017 – Jul 2019).
The City School — City School Nowshera (Sep 2016 – Dec 2017).
Alfred Sutton Primary School, Reading and New Town Primary School, Reading for earlier schooling.

Publications

Peer-reviewed research across machine learning, cybersecurity and trustworthy AI evaluation.

Governance AI Policy Institutional Design

Beyond Alignment: Governing the AI Transition through Redistributive, Institutional, and Normative Co-ordination

Submitted to Elsevier Technology in Society
View submission
Malware Cybersecurity Feature Analysis

Dissecting Detection: Feature-Class Ablation and a Deployment-Aware Evaluation Metric for Static PE Malware Classifiers

Submitted to Springer Nature Journal of Computer Virology and Hacking Techniques
View submission
Mathematics Linear Algebra Education

Geometric Intuitions for Linear Algebra: A Derivation-First Introduction for Machine Learning Practitioners

Submitted to Springer Nature Computer Science
View submission
NLP Spam Detection Classical NLP

Linguistic Deception Signatures for Cross-Corpus Spam Detection with Classical NLP

Submitted to Springer Nature International Journal of Machine Learning and Cybernetics
View submission
Safety LLMs Benchmarking

SafeCalib: Benchmarking Security-Critical Refusal Calibration in Instruction-Tuned Language Models

Submitted to Springer Nature Cybersecurity
View submission Dataset
Cybersecurity Machine Learning Benchmarking

Significance-Aware Benchmarking of Classical Machine Learning Classifiers for Phishing Website Detection

Submitted to Springer Nature Evolutionary Intelligence
View submission
Cybersecurity Transformers Tabular Benchmarks

Significance-Aware Evaluation of Classical and Transformer Classifiers Across Cybersecurity and Financial Tabular Benchmarks

Submitted to Springer Nature Evolutionary Intelligence
View submission

Projects

Selected engineering work spanning AI platforms, developer tools and SaaS systems.

Khwand

Active

AI-powered, self-healing CI/CD platform for AI-native engineering teams. Automatically detects, diagnoses, and fixes pipeline failures without human intervention using a LangGraph multi-agent orchestration system.

Next.js 14 FastAPI LangGraph Supabase Redis
View on GitHub

24syse

2024

SaaS management system for mobile laptop repair shops — integrating device history security checks, business analytics, digital record-keeping, and remote management tools for small business owners.

C# HTML/CSS SaaS
View on GitHub

Open Source

Repositories and contributions to the open-source ecosystem.

Khwand-Agent

Multi-agent orchestration framework for AI-native CI/CD pipelines

42 8

Khwand-RAG

Production-grade RAG pipeline with hybrid retrieval and re-ranking

38 5

CyberSecBench

Benchmarking suite for cybersecurity ML classifiers with statistical testing

56 12

TrustworthyLLM

Framework for evaluating LLM trustworthiness across safety and fairness

34 7

pandas

Contributed BYearEnd docstring with examples — pandas-dev/pandas

60k+ 24k+

Journey

2022

Started independent ML research

Began exploring machine learning, cybersecurity and trustworthy AI — building the foundation for future work.

2023

First publication

Published first peer-reviewed paper on significance-aware benchmarking for phishing detection classifiers.

2024

Released open-source tooling

Shipped CyberSecBench and TrustworthyLLM frameworks. Merged first contribution to pandas (pandas-dev/pandas).

2025

Published datasets

Released cissp-llmbench and safecalib-bench on Hugging Face for public experimentation and benchmark evaluation.

2026

Founded Khwand AI

Launched Khwand AI — an emerging startup focused on practical ML systems, trustworthy AI and developer tooling.

Tech Stack

Tools and technologies used in research, engineering and production systems.

PyTorch Transformers LangChain FastAPI Docker Kubernetes AWS CUDA Scikit-learn HuggingFace TensorFlow Python SQL C# Next.js Git Linux Redis Supabase LangGraph LaTeX Pandas GitHub Actions

Get In Touch

Open to research collaborations, open-source contributions, startup founding opportunities, and roles as a researcher or engineer in early-stage startups.

Email GitHub Scholar HuggingFace ORCID LinkedIn