Kianoosh Vadaei

Kianoosh Vadaei

AI Enthusiast & Developer

Computer Engineering student at the University of Isfahan passionate about AI, ML, and DL. Exploring the frontiers of artificial intelligence through research and practical applications.

View CV Email Me
15+ Projects
4 Publications
3.59 GPA

About Me

I'm a passionate Computer Engineering student at the University of Isfahan with a deep fascination for artificial intelligence, machine learning, and deep learning. My journey in AI began with curiosity about how machines can learn and make decisions, leading me to explore various domains including Natural Language Processing, Computer Vision, and AI Explainability.

Beyond academics, I enjoy reading philosophy, playing piano and dulcimer, and diving into books and video games. These diverse interests help me approach problems from unique perspectives and maintain a balanced view of technology's role in society.

NLP Computer Vision AI Explainability Interpretability Speech Processing Audio Processing

Education

B.E., Computer Engineering
University of Isfahan
September 2021 – Present
GPA: 3.59/4.0
Bachelor's Research Project
Exploring Self-Supervised Learning in NLP
Ongoing

Latest Updates

PrismSSL

A comprehensive toolbox for self-supervised learning across multiple modalities including audio, vision, graph, and cross-modal learning. Built with PyTorch and featuring advanced techniques like LoRA and distributed training.

Audio Vision Graph Cross-Modal PyTorch HuggingFace LoRA DDP WandB
GitHub Project Website

Selected Publications

From Scenes to Semantics: PersianCLEVR for Bilingual 3D Visual Reasoning

1st Workshop on VLM4RWD@ NeurIPS 2025
Co-First Author
Kianoosh Vadaei, Melika Shirian, Arshia Hemmat, Mohammad Hassan Heydari, Ali Mamanpoosh, Afsaneh Fatemi

We introduce PersianClevr, the first bilingual benchmark for 3D visual reasoning in English and Persian, built from CLEVR, Super-CLEVR, and ClevrTex with new QA synthesis. It evaluates VLMs on attributes, counting, comparison, spatial relations, and logic, revealing strong surface understanding but clear weaknesses in true 3D and compositional reasoning.

3D VQA Multimodal Benchmark Dataset: PersianClevr
OpenReview

Adaptive Chunking for VideoRAG Pipelines with a Newly Gathered Bilingual Educational Dataset

CSICC 2025 (29th International Computer Conference, Computer Society of Iran), Tehran
Co-First Author
Arshia Hemmat, Kianoosh Vadaei, Melika Shirian, Mohammad Hassan Heydari, Afsaneh Fatemi

This paper presents an adaptive chunking strategy for VideoRAG pipelines using a newly collected bilingual educational dataset. We demonstrate improved performance in video question answering through optimized temporal segmentation techniques.

VideoQA RAG Multimodal Dataset: EduViQA
IEEE Xplore

Leveraging Retrieval-Augmented Generation for Persian University Knowledge Retrieval

2024 15th International Conference on Information and Knowledge Technology (IKT)
Co-First Author
Arshia Hemmat, Kianoosh Vadaei, Mohammad Hassan Heydari, Afsaneh Fatemi

We introduce a novel approach to Persian knowledge retrieval using RAG techniques, specifically designed for university-level educational content. Our system demonstrates significant improvements in accuracy and relevance for Persian language queries.

RAG Persian LLMs Dataset: UQB
ArXiv

Skills

Advanced

Python Advanced
C++ Advanced
PyTorch Advanced

Intermediate

Keras Intermediate
C# Intermediate
Linux Intermediate
Django Intermediate

Familiar

Qt Familiar
Arduino Familiar
.NET Familiar
Flutter Familiar

Experience

AI R&D Intern

Hoopad Company
June 2024 – August 2024

Developed Siamese networks for face recognition systems. Integrated Speech-to-Text (STT), Text-to-Speech (TTS), and Large Language Model function calling capabilities into production systems.

Chairman

UIAI (University of Isfahan AI Community)
March 2023 – May 2025

Leading the university's AI community, organizing events, workshops, and research collaborations. Also serving as Chief Secretary for the Gradient magazine.

Organizer & Co-Organizer

TEDx University of Isfahan
August 2023 – Present

Organized TEDx events with 100+ participants, managing logistics, speaker coordination, and event execution. Currently serving as Co-Organizer for ongoing events.

Executive Committee Member

7th International Conference on IoT & Its Application
2023

Contributed to the organization and execution of the international IoT conference, handling technical aspects and participant coordination.

Teaching Assistant & Course Instructor

University of Isfahan & Faradars
2022 – Present

Teaching Assistant for multiple computer science courses. Course Instructor for "Fundamentals of Python" and "RAG Fundamentals" at Faradars online platform.

Curriculum Vitae

Download my complete CV to learn more about my academic background, research experience, and achievements.

Download CV

Get In Touch

Telegram
@kia_vadaei

Send a Message

This form will open your email client with a pre-filled message. No data is stored on this website.