Research portfolio · Dhaka, Bangladesh

Bengali NLP, retrieval reliability, and multimodal evaluation.

I'm Khan Raiyan Ibne Reza, a computer science graduate of North South University. I build benchmarks and AI systems to study how well language and multimodal models use evidence in low-resource settings.

Research Selected projects CV

B.Sc. CSE · NSU · 2022–2026CGPA 3.62 / 4.00

3 accepted papers

5 preprints / under review

Khan Raiyan Ibne Reza at North South University — Khan Raiyan Ibne Reza
Computer science · Bengali and low-resource AI

Research focus

Questions behind the work

Bengali and low-resource NLP

Benchmarks and methods for language technologies where data and evaluation resources remain limited.

Retrieval reliability

How query language, evidence quality, and knowledge structure shape retrieval and grounded answers.

Multimodal and efficient models

Evaluation of visual and document reasoning, alongside compact models for offline educational use.

Accepted papers

Research papers

Full publication list

Accepted

arXiv:2608.15223

TRACE-BN: Transferring Bangla-English Tutoring Behavior to a Sub-1B Offline Language Model

Bangla-English tutoring requires more than producing a correct translation: learners also need explanations of grammar differences, awareness of their likely errors, and targeted practice. We present TRACE-BN, a curriculum-guided dataset of 4,099 structured tutoring traces for Bangla-speaking learners of English at the CEFR A1–A2 level across 13 NCTB grammar units and 15 situational domains. Each trace encodes seven multi-task fields: word-level glosses, literal translation, natural translation, Bangla contrastive grammar notes, a common learner mistake, a practice question, and its answer. We transfer this multi-component tutoring behavior to Qwen3-0.6B using LoRA with 4-bit quantization (378.3 MB GGUF) for offline, on-device deployment. On 432 held-out examples, schema validity improves from 85.4% to 95.8%, chrF++ rises from 15.28 to 34.77 (+19.49 gain), and BLEU increases from 4.52 to 21.03, supported by dual automated judges and a 100-trace educator audit (4.71/5.00 overall rating).

11th IEEE Asia-Pacific Conference on Computer Science and Data Engineering (IEEE CSDE 2026)

Paper summary arXiv

Accepted

arXiv:2610.00664

PhysicsMate: A Curriculum-Grounded Bengali Benchmark for Secondary Physics QA with Small-Model Adaptation

PhysicsMate is a curriculum-grounded Bengali benchmark for secondary-school physics question answering, built from the NCTB Grade 9–10 syllabus. It contains 1,834 question-answer pairs grounded in a multi-relational knowledge graph with 1,760 nodes and 2,600 edges. Small-model adaptation across 0.6B, 1.7B, and 4B parameter scales improves closed-book accuracy by 5.5, 15.0, and 23.3 percentage points, respectively, supporting offline educational use in settings with limited connectivity and hardware.

11th IEEE Asia-Pacific Conference on Computer Science and Data Engineering (IEEE CSDE 2026)

Paper summary arXiv

Accepted

Beyond Chunking: Node Integrity and Field Structure in Bengali Retrieval-Augmented Agent Memory

This work studies how chunking and record structure affect Bengali retrieval-augmented systems and agent memory. It compares chunked text, intact full-node text, and field-structured retrieval across agricultural, legal, and secondary-education domains, examining both retrieval quality and downstream evidence grounding.

Third Workshop on KNOwledge GRaphs & Agentic Systems Interplay (NORA), AACL-IJCNLP 2026

Paper summary

Selected projects