Skip to content
View eyasir2047's full-sized avatar
🎯
Focusing
🎯
Focusing
  • Aracal

Organizations

@Android-Lab-Project

Block or report eyasir2047

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
eyasir2047/README.md

Website Google Scholar LinkedIn HuggingFace


πŸ§‘β€πŸ”¬ About Me

I'm a final-year CSE student at the University of Dhaka specializing in Large Language Models and Machine Learning. My research interests are at the intersection of AI safety, Large Language Models (LLMs) and Multimodal Systems.

  • πŸ”¬ Research Assistant @ University of Dhaka under Dr. Muhammad Ibrahim
  • πŸ“„ arXiv Author β€” NCTB-QA: 87,805 Bangla QA pairs from 50 national textbooks
  • 🌲 Section Leader @ Code in Place 2026, Stanford University
  • 🌏 Incoming NUS Young Fellow 2026 β€” National University of Singapor
  • πŸ† Champion @ Solvio AI Hackathon 2025 (650+ teams) with ZeroBin

"AI/ ML is the best thing happended to Mankind"


πŸ“„ Publications

arXiv
2026

NCTB-QA: A Large-Scale Bangla Educational Question Answering Dataset and Benchmarking Performance

Abrar Eyasir, Tahsin Ahmed, Muhammad Ibrahim

87,805 QA pairs Β· 50 NCTB textbooks Β· Benchmarks for BERT, RoBERTa, ELECTRA, Qwen-2.5-7B Β· Chain-of-Thought fine-tuning Β· BLEU / F1 / EM evaluation

Paper PDF

arXiv
2026

A Benchmark of Classical and Deep Learning Models for Agricultural Commodity Price Forecasting on A Novel Bangladeshi Market Price Dataset

Abrar Eyasir, et al.

AgriPriceBD: 1,779 daily retail prices for five Bangladeshi commodities (July 2020 – June 2025), extracted from 800 government PDFs via an LLM-assisted digitization pipeline.

Paper PDF


πŸ”­ Research Interests

Large Language Models (LLMs)          β†’  Fine-tuning, PEFT/LoRA, Chain-of-Thought, RLHF
Low-Resource NLP                       β†’  Bengali/Bangla QA, dataset construction, benchmarking
LLM Evaluation & Alignment            β†’  Empirical safety, robustness
Multimodal AI                          β†’  Vision-Language Models, cross-modal reasoning

πŸ—οΈ Featured Projects

πŸ† Champion Β· Solvio AI Hackathon 2025 (650+ teams)

End-to-end AI system for smart city waste management:

  • YOLOv8 + MobileNetV2 for real-time waste classification
  • LSTM bin fill-level forecasting Β· LightGBM e-waste pricing
  • sentiment analysis (94.5% accuracy)
  • FastAPI Β· Docker Β· Vercel deployment

Demo GitHub


πŸ› οΈ Technical Stack

πŸ€– LLMs & NLP

Transformers BERT Qwen LoRA BanglaBERT

πŸ§ͺ ML & Deep Learning

PyTorch TensorFlow scikit-learn

πŸš€ MLOps & Deployment

FastAPI Docker HuggingFace Hub Unsloth

πŸ’» Languages

Python C C++


πŸ“ˆ GitHub Stats

->

πŸ“« Get in Touch

I'm actively seeking PhD / Research positions in LLMs, VLMs, NLP and AI Safety.

Email Website Scholar


"Low-resource doesn't mean low-priority."

Popular repositories Loading

  1. FoodiePal FoodiePal Public

    Java 1

  2. Sirius-Academy-Slides Sirius-Academy-Slides Public

    This repo contains higher math slides of Sirius Academy

    1

  3. eyasir2047 eyasir2047 Public

    Config files for my GitHub profile.

  4. C-function-implementation C-function-implementation Public

    Java

  5. Data-Structure-Algorithm Data-Structure-Algorithm Public

    C++

  6. LeetCode-Solution LeetCode-Solution Public

    LeetCode DSA

    C++