LLM Engineering & Evaluation
Build language systems you can measure
The Institute of Artificial Intelligence · Lahore
Rigorous programs, applied research and an AI-native campus — where engineers, researchers and leaders learn to build systems people can trust.
Three ways in
Whether you are here to learn, to research, or to build capability across an organisation, the institute is designed around the work you will do.
Diplomas, certificates and executive education taught by faculty who still ship. Every program ends in defended work.
9 programs · 12 courses
See programsFive labs working on language, perception, safety, systems and health — with a bias toward problems that matter here.
10 publications · 4 active projects
Visit the labsCohorts, private tracks and a portal that shows managers exactly how their teams are progressing.
Seats, reports and invite codes
Partner with usPrograms
From a first course in machine learning to a funded research year, each program is cohort-based, project-heavy and reviewed by faculty and industry examiners.
Nine months from working engineer to AI practitioner who ships.
A rigorous first course for people who want to understand, not just use.
Build language systems that survive contact with real users.
Six weeks for the people who decide what gets built.
Course catalog
Every course runs on the campus: lessons, quizzes, browser-based labs and a tutor that has read the syllabus.
LLM Engineering & Evaluation
Build language systems you can measure
AI Safety, Alignment & Governance
What could go wrong, and what to do about it
Foundations of Machine Learning
From first principles to a model you can defend
The campus
The course player, coding lab and tutor share one context. The tutor knows which lesson you are on and can read the code in your lab — and it is tuned to teach, not to solve.
Socratic tutor
Asks before it answers. Grounded in the lesson, the brief and your work.
Browser-based labs
Python and JavaScript run locally in your browser with tests, no setup.
Progress that means something
Quizzes, labs and projects roll up into verifiable certificates.
Research
Urdu-language models, vision on sub-$50 hardware, evaluation standards for public deployment. Our labs work on what benchmarks ignore.
Fatima Noor, Hamza Qureshi, Sana Malik · PIOAI Policy Report 2026-01
Bilal Ahmed, Omar Farooq, Hamid Raza · ICLR 2026
Bilal Ahmed, Rafia Hussain, Ali Hassan, Sana Malik · ACL 2026 (Findings)
Journal
Static benchmarks decay the moment they enter training data. Here is how we built one that stays honest, and what it tells us about the state of Urdu language technology.
Forty places, two scholarships for women in engineering, and a new capstone partnership with the provincial health department.
What we learned deploying crop-disease detection on sub-$50 hardware across forty-one farms in Punjab — including everything that broke.
Admissions
Forty places on the Professional Diploma, two full scholarships for women in engineering, and a rolling intake for Foundations. The application takes about forty minutes.