A selection of senior projects and MS theses supervised in the lab, grouped by research theme. Work that reached publication is marked as such; work that stopped at a tested prototype or a field study is listed with what it actually found. Both are part of the record.
A culturally-grounded explainable AI system for autism developmental screening in South Asia, correcting for parental scoring bias. Action-imitation and voice-imitation tasks each paired with an explainability layer so clinicians can see why a score was given.
A school-deployable ADHD screening pipeline for Pakistani children aged 8–12, since existing objective tools require fMRI or EEG hardware unavailable locally. Dus Pattay is a Unity tablet game across five measurement domains; Khoj is a bilingual parent-education companion.
How social robots can support speech-language therapy for Pakistani children, across ASD-related communication deficits, articulation, developmental language disorder, stuttering, dysarthria and bilingual code-switching gaps.
A probabilistic Urdu-language dyslexia screening tool. An estimated 15–20% of Pakistani children show signs of dyslexia, but no indigenous-language assessment tool existed. Eight tests across three grade bands, with Urdu word data collected from special-education institutes.
An eight-month case study with six diagnosed dyslexic students and two psychologist-teachers identifying the specific error patterns Urdu writers with dyslexia make — phonological confusions, join-word misspellings, skipped short vowels, letter-shape confusions — compared against Arabic, Persian and English.
Gameplay and assessment data from 680 primary-school children across six Lahore schools. Gaussian Mixture Model clustering uncovered latent risk subgroups, then a Random Forest classifier stratified severity.
Three Unity mini-games targeting the reading sub-skills Urdu-speaking children with dyslexia struggle with most, grounded in the Dual Route Model and Urdu's dual orthography. Tested with 15 children at an SOS Village orphanage in Karachi.
Five tablet-based learning applications plus an adapted version of Guess Who?, field-tested with eight autistic children over seven weeks at a specialist school in Lahore, targeting turn-taking, question-asking, categorisation and social attention.
A digital toolkit for aphasia, developed with neuro-physicians and speech-language pathologists in Lahore. Three prototypes explored different assistive pathways: OCR-based handwriting recognition, speech-to-text detection, and a VR interaction environment.
Whether low-cost computer vision could screen for autism-related behavioural markers, as an alternative to lengthy specialist-dependent clinical assessment. Reviewed eye-tracking, motor-skill and affective-computing approaches.
A platform for university counselling centres built on React, Supabase and FastAPI. Its Acoustic-Text Emotional Incongruence engine compares what a client says against how they say it — jitter, shimmer, vocal intensity and speech rate against a RoBERTa linguistic model — surfacing moments where words and vocal affect do not match.
Gamified CBT delivered inside a to-scale 3D recreation of the students' own school, with scenarios embedded at real campus locations. A parallel 2D prototype served as the control condition.
A mental health support solution for Pakistani high-school students addressing academic stress, isolation and distrust of available resources. A 61-respondent survey, school interviews at three campuses, and a co-design workshop with scenario cards.
Natural virtual environments with mindfulness and breathing exercises, grounded in exposure therapy. The team pivoted from an indoor therapy room to a nature setting after expert feedback that nature environments outperform closed spaces.
Whether interactive storytelling can encourage Pakistani high-schoolers to engage with their mental health, given high stigma and low access to therapy. Storyboarding and interviews surfaced grade anxiety, favouritism, harassment and self-harm.
An ethnographic study of 17 LUMS undergraduates: two weeks of daily app-usage logging, screen-time capture around deadlines, four interviews and a focus group.
Staged behaviour-change journeys across sleep, focus, fitness and mood, grounded in the Fogg Behavior Model and the Transtheoretical Model.
An avatar-driven branching narrative supporting students around social anxiety and depression, set in orientation week. Grounded in avatar-based digital therapeutics, serious games and CBT for social anxiety.
A predictive risk-modelling system aimed at identifying students at elevated risk before a crisis develops, in the same computational mental-health stream as CareSync.
An adaptive app diagnosing individual students' specific addition mistakes, against ASER data showing only ~51% of rural grade-5 students can do grade-2-level math. A pilot at a government school, then a main study on addition with 220 grade-3 students.
Addressing the decline of creative Urdu writing in Pakistani primary schools with a gamified storytelling app: an Urdu synonym dictionary, a doodle pad, a curated image bank and Urdu speech-to-text dictation.
An education-focused project in the research and discovery phase, with a literature review and a second round of user interviews complete.
Whether gamified VR simulating everyday activities improves engagement and outcomes in upper-limb rehabilitation for stroke, Parkinson's and spinal injury patients. A Unity XR virtual house with five tasks targeting wrist rotation, reach and hand-eye coordination.
A culturally-grounded system for informal Pakistani family caregivers, addressing both logistical burden and emotional stressors. Interviews identified six themes: emotional labour, medical management, social isolation, filial piety, elder resistance to help, and stress buffers.
A telerehabilitation platform for PSRD Hospital, letting lower-income physically disabled patients access physiotherapy remotely instead of facing financial, time and weather barriers to in-person visits. Doctor, patient and admin applications with a bilingual interface.
How Pakistanis aged 50 and older actually use, or avoid, mobile phones — 102 surveyed users split between smartphone and feature-phone owners, with 20 follow-up interviews.
Two phases: immersive VR nature environments for relaxation, then a VR physiotherapy exercise tool for patients of the Pakistan Society for Rehabilitation of the Disabled.
An early-stage dementia and ageing technology project.
A bilingual, AI-mediated web app challenging implicit gender stereotypes among Pakistani pre-teens aged 9–12. Formative research with 24 children and 20 parents found boys assigned to leadership roles and girls to caregiving. Children drag characters into roles across a map of everyday locations, then engage in Socratic LLM dialogue that prompts reflection without correcting them.
A safe space for Pakistani women aged 15–40 to find help after harassment, combining legal information, emotional support and reporting tools. Built on 136 survey responses and interviews with a lawyer and a harassment-committee member.
Family planning information and cycle tracking for Pakistani women, addressing the lack of culturally appropriate, private resources. A bilingual prototype with illustrated stories, a contraceptive-guidance chatbot, a disguised private mode, and content encouraging male involvement.
How manipulative UX affects children in Pakistan, and whether socioeconomic status and parental supervision change vulnerability. A drawing and survey study with 16 grade-5 children mapping their most-used apps.
Helping mothers track communication, motor and personal-social milestones from 2 to 36 months, alongside awareness content and activities. User research with 19 mothers and doctor interviews.
Why parental-control apps popular in the West see low adoption in Pakistan. 84 parents surveyed, 33 existing apps benchmarked, and two shortlisted apps usability-tested with 30 parents.
A narrative and illustration-based intervention raising awareness of sexual harassment on campus and encouraging bystander intervention, grounded in LUMS harassment-policy categories.
A virtual tour of the historic Walled City of Lahore, since much of its heritage is unmarked or mislabelled on Google Maps. Students walked the walled city, verified map locations and compiled structured data and photo documentation for each historic gate and surrounding neighbourhood.
An ARCore Android application testing whether AR guides children through museum exhibits more effectively than a conventional guided tour, overlaying 3D models, animations and contextual information onto real exhibits.
A digital guide to the Rafi Peer Cultural Centre puppet collection, with an events calendar, directions and an object-scan feature.
A VR firefighter-training system developed with the Punjab Emergency Service (Rescue 1122) Academy, teaching cognitive skills like scene size-up that are hard to practise repeatedly in physical simulators. Seven expert interviews and a focus group with instructors.
A Unity VR project on distracted and stressed driving scenarios, aimed at emotion detection and regulation for road-safety awareness.
A free web app applying computer vision to creative photo editing: photomosaic generation, neural style transfer and a crayon-effect filter using bilateral filtering and edge detection.
Most of this work started as a course project or a senior project. Students who want to take one further join the lab as research assistants and continue towards publication.