
Digitize Bilingual Curricula
Transform the Auto-Electrician curriculum into Urdu and English, enriched with interactive simulations.
Upscaling & Digitisation of Vocational Education in Pakistan
The Ba-Ikhtiyar Jawan project is designed to transform Pakistan’s Technical and Vocational Education and Training (TVET) sector by embedding artificial intelligence, multimodal learning, and international collaboration.
As a proof of concept, the project focuses on digitizing and upscaling the Auto-Electrician trade, a profession with high demand in Pakistan and across Gulf labor markets. By leveraging AI-powered language technologies and adaptive learning pathways, the initiative aims to make vocational education more inclusive, accessible, and industry-aligned.
Our initiative combines cutting-edge AI, global collaboration, and practical skill-building to create a scalable, future-ready training ecosystem.
Urdu/English course outlines enriched with interactive simulations, sourced from CTTIs, TEVTA, and German partners — bridging theory and hands-on practice.
From an Urdu Small Language Model to German–English machine translation and a personalized recommender system, our AI backbone adapts learning to each individual.
Text, voice, diagrams, and video lectures — including content from CTTI, German experts, and local innovators — to strengthen practical skills.
Optimized for rural and resource-constrained environments, ensuring no learner is left behind.
Partnerships with institutions in Germany, UAE, and Saudi Arabia to align training with international standards.
Our initiative combines cutting-edge AI, global collaboration, and practical skill-building to create a scalable, future-ready training ecosystem.
Digitize Bilingual Curricula
Transform the Auto-Electrician curriculum into Urdu and English, enriched with interactive simulations.
Develop Urdu & English SLMs
Create domain-specific Small Language Models to power adaptive, interactive conceptual learning.
Recognize Prior Skills
Enable formal recognition for workers transitioning from informal or on-the-job training.
German-to-English MT System
Translate technical manuals from German to English for wider accessibility.
Personalized Recommendation Engine
Deliver equitable, tailored learning pathways for every learner.
Multimodal Learning Modules
Integrate text, voice, diagrams, and visual cues to strengthen hands-on skills.
Scalable Multi-Trade Digitization
Build capacity to digitize multiple vocational trades for long-term sustainability.
Our work creates value across the entire vocational training ecosystem — from learners to policymakers.
Auto‑Electrician trainees, especially from marginalized communities, gain access to modern, bilingual, and practical training.
Curriculum designers and instructors receive standardized, digitized resources and adaptive teaching tools.
Employers benefit from a steady pipeline of skilled, job‑ready technicians.
Gulf and overseas markets gain access to certified, globally aligned talent.
NAVTTC, NEVTCC, and other authorities can align standards, track outcomes, and scale best practices.
Our work spans two interconnected dimensions — advancing cutting‑edge technology and fostering collaboration to ensure long‑term impact.
Domain‑adapted AI to support conceptual learning in Auto‑Electrician training.
Tailored for translating technical manuals.
Personalized learning pathways for every student.
Simulations and interactive content using text, voice, diagrams, and visual cues.
Lightweight systems for resource‑constrained areas.
Workshops and training to reduce resistance and enable adoption of digital curricula.
Curriculum consultations with domestic employers and Gulf labor markets.
Knowledge‑sharing, joint reviews, and exchange visits with institutions in Germany, UAE, and Saudi Arabia.
We’ve divided the project into five focused work packages, each with clear objectives, deliverables, and timelines to ensure measurable progress.
Objective: Modernize the Auto‑Electrician curriculum with bilingual (Urdu/English) modules.
Activities: Develop interactive modules, digitize content.
Deliverable: Digitized curriculum package.
Milestones: Beta in 6 months, final in 12 months.
Objective: Build domain‑specific SLMs for conceptual learning.
Activities: Collect data, train models, integrate into platform.
Deliverable: Functional SLM for adaptive training.
Milestones: Prototype in 9 months, full integration in 12 months.
Objective: Create practice‑oriented, multimodal training tools.
Activities: Simulations, interactive modules, multimodal interaction.
Deliverable: Practical training modules.
Milestones: Pilot in 12 months, industry‑validated in Year 2.
Objective: Upskill instructors and align curricula with market needs.
Activities: Workshops, training sessions, industry consultations.
Deliverable: Trained instructors, updated curriculum.
Milestones: 3 workshops in Year 1, updates by Month 12.
Objective: Leverage expertise from Germany & Gulf partners.
Activities: Exchange visits, joint curriculum reviews, best‑practice sharing.
Deliverable: Internationally aligned curriculum.
Milestones: 2 exchange visits, review by Month 10.
We follow a structured, iterative process from data collection to evaluation to ensure innovation, accuracy, and real‑world impact.
Data Collection
📚 Gather German Auto‑Electrician manuals and Urdu/English training content.
🛠 Preprocess, clean, and annotate datasets for multilingual model training.
Machine Translation (MT)
🌐 Build a German→English MT model using open‑source frameworks.
🎯 Fine‑tune for technical vocabulary; evaluate with BLEU, TER, and human review
Urdu Small Language Model (SLM)
🤖 Develop a domain‑specific Urdu SLM for contextual learning, Q&A, and adaptive assessments.
⚡Optimize for edge deployment with voice input/output.
Recommendation Engine
🎯 Personalize learning pathways using learner profiles, progress data, and feedback loops.
📈 Continuously measure engagement and skill acquisition.
Multimodal Learning
🎥 Create interactive, practice‑based modules with text, speech, diagrams, and visual cues.
🌍 Ensure accessibility for diverse literacy and technical backgrounds.
System Integration & Edge Deployment
🔗 Combine MT, SLM, Recommendation Engine, and multimodal modules into one platform.
📦 Deploy on lightweight, offline‑capable devices; pilot test with trainees.
Evaluation
📊 Assess translation accuracy, learning effectiveness, and usability.
🔄 Refine based on pilot feedback to enhance outcomes.
Our initiative delivers tangible, scalable results — transforming vocational training for Pakistan’s youth and aligning with global standards.
Interactive Urdu/English modules for Auto-Electrician training.
Adaptive, domain-specific AI for conceptual learning.
Practice-based training using text, voice, diagrams, and visual cues.
Accurate translation of technical manuals for wider accessibility.
Personalized learning pathways tailored to each learner.
Upskilled educators equipped with digital and globally aligned practices.
Joint curriculum development with partners in Germany, UAE, and Saudi Arabia.
Extendable to other vocational trades beyond Auto-Electrician.
Machine Vision and Intelligent Systems Lab | NUST-SEECS | © 2025