Open to opportunities

Abdurrehman Tariq

@abdurrehmantariq1

Message

I build efficient deep learning systems for speech and multilingual language analysis.

Pakistan

Message

What I'm looking for

I'm looking to apply my work in deep learning, speech processing, NLP, and model optimization to build measurable, efficient AI systems while continuing to grow through challenging research and engineering work.

I've built LightSER Optimized, a parameter-efficient speech emotion recognition model accepted at IBCAST 2026, achieving 96.75% accuracy and 96.76% F1-score with over 33× fewer parameters than the reproduced baseline.

I'm a Data Science undergraduate at FAST NUCES focused on deep learning for speech and language tasks. I also developed a multilingual sentiment analysis system using XLM-T and prompt-based fine-tuning, improving accuracy by up to 1.84% across languages and accelerating robustness testing by 1.89×.

At COMSATS Internet Services, I researched and presented technical recommendations for server-room temperature monitoring and internal ticketing R&D initiatives.

Experience

Work history, roles, and key accomplishments

Current

Management Trainee

Current

COMSATS Internet Services

Jun 2025 - Present (1 year 3 months)

Researched and presented technical recommendations for two R&D initiatives, a server room temperature monitoring system and an internal ticketing system, to the Senior Manager Software, informing subsequent development decisions.

Research And Presentations

Education

Degrees, certifications, and relevant coursework

FAST NUCES

Bachelor of Science, Data Science

2023 -

Grade: 3.46

Pursuing a Bachelor of Science in Data Science with a GPA of 3.46, covering object-oriented programming, data structures, algorithms, and statistical analysis.