Ammar Faisal

Ammar Faisal

Rank 1 CS @ LUMS | AI Researcher (LLMs & Security) | President, LUMS Religious Society

Based in: Lahore, PakistanCurrently: Leader, Peer Ambassadors for Social Support (PALss)Experience: 2 years

Hire me

Top-ranked Computer Science student at LUMS (Rank 1, Class of 2027) and Manzurul Haq Scholar. Prolific researcher in LLMs, multi-agentic systems, and data science, with work appearing on ClinicalTrials.gov and manuscripts in progress for Nature Health. Proven leader as President of the LUMS Religious Society, where I manage 200+ members and build custom internal tools using Next.js and Supabase to digitize operations.

Experience

Leader

Peer Ambassadors for Social Support (PALss)Lahore, Pakistanvolunteer2025 – Present

Summary

Undergraduate Advising Office, School of Science and Engineering, LUMS.

What I did

  • Promoted to PALss Leader following strong peer reviews after a year as a PALss supporting the social and mental well-being of 14 assigned first-year students.
  • Currently overseeing 19 PALss, who collectively support the incoming first-year cohort within the school.

Research Assistant

Distributed and AI Systems Lab @ LUMSLahore, PakistanworkSep 2026

Summary

Working on two early-stage projects: (1) child safety risks in large language models, and (2) security vulnerabilities in multi-agentic systems.

Results

  • Discovered that common single jailbreak strategies remain effective in multi-turn attacks, while older methods like 'DAN' and 'Override Instructions' show diminished performance.
Large Language ModelsMulti-agentic systemsSecurityJailbreakingMulti-turn attacks

Research Assistant

Internet, Data, and Society Lab @ LUMSLahore, PakistanworkMar 2025

Summary

Effectiveness of AI-Assisted Antibiotic Prescribing. In collaboration with Dr. Danesh Moradigaravand (KAUST). Misinformation Intervention Study. Generative AI Literacy Assessment Tool.

What I did

  • Designed a 10-point composite instrument (APAS) spanning four clinical domains, evaluated through 12 proctored clinical vignettes; drafted and registered a pre-analysis plan; managing participant recruitment and classification.
  • Designed and conducted randomized control trial with 250 participants, developing AI-generated podcast intervention using NotebookLM and evaluating effectiveness through mixed-effects modeling (pymer4) with comprehensive subgroup analyses across age, education, political affiliation, and media consumption patterns.
  • Leading development of a validated GenAI literacy instrument for general population, mapping 30 items to UNESCO AI Competency Framework across four key dimensions and 3 progression levels.

Results

  • Achieved a 22% reduction in sharing intentions for false news through an AI-generated podcast intervention.
  • Protocol registered on ClinicalTrials.gov; data collection ongoing.
  • Found significant reduction in sharing intentions for all news types and improved classification accuracy for false news across intervention group, with differential effects by participant demographics.
  • Data collection and analysis complete; currently authoring manuscript for Nature Health.
AINotebookLMpymer4Mixed-effects modelingGenerative AIUNESCO AI Competency Framework

Skills

Technical

Generative AI
Data Science
Large Language Models
SupabaseSupabase
Machine Learning
Next.jsNext.js
STATA
pymer4
Distributed Systems
VercelVercel

Areas of expertise

Teaching
Leadership

Projects

LRS Application Shortlisting & Interview Portal

Developerpersonal
Built a shortlisting and interview management console for LRS General Body applications, used by directors and the EC to review candidates and manage interview scheduling.
Next.jsSupabaseVercel

LRS Task Management Platform

Developerpersonal
Built a role-based internal management platform for LRS enabling task assignment, historical tracking, and meeting scheduling across the EC, directors, and assistant directors.
Next.jsSupabaseVercel

Education

B.S. in Computer Science

Lahore University of Management SciencesComputer ScienceGPA 4Sep 2023 – Jul 2027

Coursework

Distributed SystemsFoundations of AI and Machine LearningFoundations of Generative AITopics in Large Language Models (graduate-level)Internet of ThingsIntroduction to Blockchain TechnologyEconometrics

Societies & activities

LUMS Religious Society (LRS)Peer Ambassadors for Social Support (PALss)

Recognition

Awards

Rank 1, BS Computer Science, Class of 2027

Lahore University of Management Sciences2026

Ranked first in the BS graduating class of 2027 at the end of the third year.

2nd Place, LUMS STATA Competition

Chaudhry Nazar Muhammad Department of Economics, LUMS2026

Top Ranked Student (2023-2025)

School of Science and Engineering, LUMS2025

For both the first, second and third academic years.

Manzurul Haq Scholar (2024-2025)

Lahore University of Management Sciences2025

Awarded for securing top position across the entire LUMS cohort (1000+ students from all disciplines) at the end of the second year.

Patents & publications

Effectiveness of AI-Assisted Antibiotic Prescribing

publication

Target Journal: Nature Health. Protocol registered on ClinicalTrials.gov.

With Dr. Ihsan Ayyub Qazi, Dr. Danesh Moradigaravand

Generative AI Literacy Assessment Tool

publication

Currently authoring manuscript. Mapping 30 items to UNESCO AI Competency Framework.

Governance & advisory

President

LUMS Religious Society (LRS)other2025 – Present

Leading a society of approximately 200 members, overseeing strategy, events, and directorate operations.

Hire AmmarReach out about a role, a contract or a conversation.For recruiters

Ammar's twin is AI, it can make mistakes.

More people on Kanvis

Browse more peopleCreate your own Kanvis