
K. M. Jubair Sami
I'm an AI engineer at Link3 Technologies in Dhaka, Bangladesh's largest ISP, where I build LLM agents and agent harnesses, and the data pipelines and production systems around them.
My research looks at how LLMs treat low-resource languages and their dialects, starting with Bengali. My first-author work on Bengali dialectal bias appears at EMNLP 2026 (main conference) and the BLP-2025 workshop. It grew out of my undergraduate thesis at BRAC University, which won the Best Thesis Award. I graduated in Computer Science.
Outside work I build MappedTech, a laptop comparison platform with an AI agent that cites its sources. I've made tech videos on YouTube since 2017.
Open to research collaborations and graduate positions in low-resource and dialectal NLP, LLM evaluation, and agentic AI and agent security.
Try
Updates
- Paper published at EMNLP 2026, Main Conference
- Session Chair, EMNLP 2026
- Started ChurnSync
- AI Engineer (Full-time) at Link3 Technologies Ltd.
- Started Intelligent Ticketing System
Research
Google Scholar ↗Low-resource and dialectal NLP · Agentic AI and agent security
Benchmarking Bengali Dialectal Bias: A Multi-Stage Framework Integrating RAG-Based Translation and Human-Augmented RLAIF
A validated way to evaluate translation into unstandardized dialects: an LLM judge that correlates with humans better than legacy metrics.
ACL AnthologyPDFarXivDatasetAbstract
Large language models (LLMs) frequently exhibit performance biases against regional dialects of low-resource languages. However, frameworks to quantify these disparities remain scarce. We propose a two-phase framework to evaluate dialectal bias, operationalized as comprehension degradation relative to standard Bengali, in LLM question-answering across nine Bengali dialects. First, we translate and gold-label standard Bengali questions into dialectal variants adopting a retrieval-augmented generation (RAG) pipeline to prepare 4,000 question sets. Since traditional translation quality evaluation metrics fail on unstandardized dialects, we evaluate fidelity using an LLM-as-a-judge, which human correlation confirms outperforms legacy metrics. Second, we benchmark 19 LLMs across these gold-labeled sets, running 68,395 RLAIF evaluations validated through multi-judge agreement and human fallback. Our findings reveal severe performance drops linked to linguistic divergence. For instance, responses to the highly divergent Chittagong dialect score 5.44/10, compared to 7.68/10 for Tangail. Furthermore, increased model scale does not consistently mitigate this bias. We contribute a validated translation quality evaluation method, a rigorous benchmark dataset, and a Critical Bias Sensitivity (CBS) metric for safety-critical applications.
A Comparative Analysis of Retrieval-Augmented Generation Techniques for Bengali Standard-to-Dialect Machine Translation Using LLMs
Standard Bengali → 6 regional dialects with very little data and no fine-tuning.
ACL AnthologyPDFAbstract
Translating from a standard language to its regional dialects is a significant NLP challenge due to scarce data and linguistic variation, a problem prominent in the Bengali language. This paper proposes and compares two novel RAG pipelines for standard-to-dialectal Bengali translation. The first, a Transcript-Based Pipeline, uses large dialect sentence contexts from audio transcripts. The second, a more effective Standardized Sentence-Pairs Pipeline, utilizes structured local_dialect:standard_bengali sentence pairs. We evaluated both pipelines across six Bengali dialects and multiple LLMs using BLEU, ChrF, WER, and BERTScore. Our findings show that the sentence-pair pipeline consistently outperforms the transcript-based one, reducing Word Error Rate (WER) from 76% to 55% for the Chittagong dialect. Critically, this RAG approach enables smaller models (e.g., Llama-3.1-8B) to outperform much larger models (e.g., GPT-OSS-120B), demonstrating that a well-designed retrieval strategy can be more crucial than model size. This work contributes an effective, fine-tuning-free solution for low-resource dialect translation, offering a practical blueprint for preserving linguistic diversity.
Work & education
- Feb 2026 – Present
AI Engineer · Link3 Technologies Ltd.
Joined as a QA engineer, moved to AI within about two weeks, and became a full-time AI engineer in June 2026. Product owner on 6 internal products for the CeX, HR and CLM departments.
ChurnSync, Intelligent Ticketing System, Call Center Automation, HR Policy AI Assistant, Observability at Link3, Link3 Visitor Portal, Link3 Onboard Gateway, Campaign Call Tracker
- Dec 2023 – Present
Educational Content Creator · MappedAcademy
Educational content on data structures and algorithms for local students.
- Feb 2025 – May 2025
Student Tutor, CSE221 Algorithms · BRAC University
Undergraduate teaching assistant for the Algorithms course for one semester.
- Jan 2017 – 2024
Tech Content Creator · YouTube · MappedTech
Tech reviews, and later laptop buying guides for Bangladesh. The channel started as a way to get over stage fright; the brand continues as MappedTech.
- Oct 2022 – May 2026
B.Sc. in Computer Science · BRAC University
CGPA 3.95 / 4.00 · Best Thesis Award (Fall 2025)
Selected projects
All 10 →Videos
@MappedTech · @MappedAcademyComplete guidelines to buy a new laptop in Bangladesh
How I kept my MacBook's battery health at 100% after 1.5 years