Professional Experience

Belsons Technologies logo

Full Stack Developer

Aug 2025 - Present
Belsons Technologies · Dubai, UAE
  • Extended a proprietary ERP system by developing and enhancing full-stack features to support evolving business needs.
  • Architected and implemented a project-wide access control system to manage roles and permissions securely.
  • Implemented interfaces for multi-tenant management and fine-grained, role-based access control across the application.
  • Created automated tests to validate core business logic and improve system reliability.
Belsons Technologies logo

Software Engineering Intern

Jul 2024 - Aug 2024
Belsons Technologies · Dubai, UAE
  • Built automated market reports using aggregate statistics and GPT-4 to generate narrative summaries.
  • Implemented a property image classifier to categorize room types for listings.
  • Developed Django admin dashboards to manage data inputs and review AI-generated insights.
  • Computed area-based aggregates (e.g., average rent, price per sqft) to support AI-generated outputs.
ILM UX Ltd. logo

Software Engineering Intern

Jun 2022 - Sep 2022
ILM UX Ltd. · Mumbai, India
  • Wrote unit tests for frontend and backend components using Jest, improving baseline reliability.
  • Built a real-time internal chat application using Socket.IO and a Node.js backend with React/Next.js.
  • Integrated AWS S3 to support image and file uploads in the chat with signed URL handling.
  • Implemented basic read receipts and message delivery statuses using WebSocket events.

Projects

Generation of High-Level Music Structures

American University of Sharjah - COE 476

  • Processed music dataset into music-prompt pairs with extracted structural information.
  • Finetuned the MusicGen transformer using a LoRA adaptor.

CLAP LAION FAD: 0.9959; Overall Quality: 70.48/100

Waste Classification with Feature Extraction

American University of Sharjah - CMP 466

  • Processed a 12-class garbage classification dataset using CNN feature extraction with TensorFlow and scikit-learn.

MobileNetV2 Accuracy: 0.93; Weighted F1: 0.93

Bank Account Fraud Detection

American University of Sharjah - STA 401

  • Used balanced data mining approaches including LDA, SMOTE, and Balanced Random Forest.
  • Achieved competitive results on a highly imbalanced dataset.

Recall: 0.79; AUC: 0.88

Attendance & Event App

AUS College of Engineering

  • Utilized Flutter for cross-platform mobile application development.
  • Integrated Node.js backend, configured notifications and offline caching.

Full-Stack Reddit Clone

Personal Project

  • Used React and Next.js to build a front-end clone of Reddit.
  • Created a Node.js + Express backend with PostgreSQL and JWT authentication.

Restaurant Management Software

American University of Sharjah - CMP320

  • Built a restaurant management system using React, Next.js, and Django.
  • Implemented user authentication and management features.

Research

Generative AI for Early Grade Story Generation Using a Self-Reflective Approach

Taufiq Syed, Aadhith Shankarnarayanan, Yara Kaddoura, Salsabeel Shapsough, Imran Zualkernan, Ekaterina Kochmar
American University of SharjahIEEE International Conference on Advanced Learning Technologies (ICALT)July 14-17, 2025
This paper presents a self-reflective story generation framework using GPT-4o to support large-scale early grade reading assessments. By grounding generation in classic tales and incorporating iterati...
View Publication DOI: 10.1109/ICALT64023.2025.00058

Tailored Tales: Enhancing Children's Reading Comprehension with Preference-Tuned Automatic Story Generation

Aadhith Shankarnarayanan, Taufiq Syed, Salsabeel Shapsough, Imran Zualkernan, Ekaterina Kochmar
American University of Sharjah11th International Conference on Computing and Artificial Intelligence (ICCAI)March 28-31, 2025
This study investigates preference tuning as a cost-effective alternative to full fine-tuning for generating high-quality children’s reading comprehension stories. Using annotated examples inspired by...
View Publication DOI: 10.1109/ICCAI66501.2025.00057

Once Upon a GPT-4: Enhancing Diversity in Automated Reading Comprehension Story Generation with Classic Tales

Aadhith Shankarnarayanan, Taufiq Syed, Salsabeel Shapsough, Imran Zualkarnan
American University of SharjahIEEE International Conference on Advanced Learning Technologies (ICALT)July 1-4, 2024
This paper addresses the challenge of generating reading comprehension content for early grade assessments—an expensive and time-consuming task. It proposes leveraging GPT-4 mediated by a classic tale...
View Publication DOI: 10.1109/ICALT61570.2024.00063

Can GPT-4 Aid in Detecting Ambiguities, Inconsistencies, and Incompleteness in Requirements Analysis? A Comprehensive Case Study

Taslim Mahbub, Dana Dghaym, Aadhith Shankarnarayanan, Taufiq Syed, Salsabeel Shapsough, Imran Zualkarnan
American University of SharjahIEEE Access (Volume 12)Published: September 19, 2024
This paper explores GPT-4's performance in identifying defects in a real-world software requirements specification for a mechanical lung ventilator. Using a zero-shot setting, GPT-4 was evaluated for ...
View Publication DOI: 10.1109/ACCESS.2024.3464242

About Me

Taufiq Syed

I’m a recent Computer Science graduate from the American University of Sharjah, with a strong foundation in AI, data science, and software development.

Alongside my studies, I’ve been actively involved in student leadership and tech communities, creating spaces that inspire innovation and collaboration.

Download Resume

Education

Bachelor of Science in Computer Science from the American University of Sharjah,
graduated in June 2025 with a GPA of 3.41.
Courses: Neural Networks, AI, Data Mining, etc.

Achievements

2nd place in SoftwareAG GITEX Future Disruptors (2023)
1st in AUS Programming Contest (2022)
2nd in ADU STEM Programming Contest (2022)

Extracurricular Activities

President - Open Source Club (2023-2024)
Activities Coordinator - Technopreneurship Club
Organized 15+ events and led a 5-part Flutter workshop series.