Skip to content
View importrayhan's full-sized avatar
:shipit:
Building Conversational Query Understanding Benchmarks
:shipit:
Building Conversational Query Understanding Benchmarks

Block or report importrayhan

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
importrayhan/README.md
Typing SVG

🔬 Research Engineer | 🤖 NLP Specialist | 🚀 LLM Enthusiast

Building intelligent retrieval systems and conversational AI that bridge the gap between human language and machine understanding


👨‍💻 About Me

I'm a Computing graduate specializing in Natural Language Processing, Information Retrieval, and Large Language Models. My research focuses on making search systems more intelligent through conversational AI and ambiguity detection.

  • 🎓 MSc Computing from Edinburgh Napier University
  • 🔍 Research: Clarification Need Prediction in Multi-turn Conversational Search
  • 📝 Published in Computers and Education: Artificial Intelligence (Q1)
  • 📄 Two papers submitted to SIGIR 2026 on adversarially robust LLMs
  • 🌱 Currently exploring RAG architectures and production LLM deployment
  • 💼 Former Research Associate building ML systems for education
  • 🌍 Based in Edinburgh, UK

🛠️ Technical Stack

Core AI/ML

PyTorch HuggingFace scikit-learn TensorFlow Pandas NumPy

NLP & LLMs

Transformers LangChain BERT GPT

GPU & Compute

CUDA Slurm AWS

Development

Python Linux Bash Git Docker Flask


🔬 Research Focus

research_interests = {
    "Information Retrieval": [
        "Conversational Search",
        "Query Understanding",
        "Ambiguity Detection",
        "Clarification Generation"
    ],
    "Large Language Models": [
        "Fine-tuning & Prompt Engineering",
        "Adversarial Robustness",
        "Retrieval-Augmented Generation (RAG)",
        "Model Evaluation & Benchmarking"
    ],
    "NLP Applications": [
        "Multi-turn Dialogue Systems",
        "Search Intent Classification",
        "Semantic Search",
        "Document Understanding"
    ],
    "ML Systems": [
        "Production ML Pipelines",
        "Model Deployment (AWS, Flask)",
        "GPU-Accelerated Training",
        "Distributed Computing (Slurm)"
    ]
}

📊 Current Projects

🔍 RAG-Enhanced Conversational Search

Building retrieval-augmented generation systems for multi-turn query clarification

  • Tech: PyTorch, Transformers, FAISS, LangChain
  • Focus: Combining dense retrieval with LLM generation

🤖 Adversarial LLM Robustness

Evaluating and improving LLM reliability under adversarial inputs

  • Tech: PyTorch, HuggingFace, Custom datasets
  • Status: Under review at SIGIR 2026

🎯 GPU-Accelerated Model Training

Fine-tuning transformer models on HPC clusters

  • Tech: PyTorch, CUDA, Slurm
  • Scale: Multi-GPU distributed training

📚 Educational AI Systems

ML pipelines for student performance prediction

  • Tech: scikit-learn, Flask, AWS EC2
  • Publication: Q1 journal (4 citations)

📈 GitHub Stats

GitHub Streak

🏆 Key Achievements

  • 📝 Published Research: Q1 journal in AI & Education (cited 4 times)
  • 🎓 Academic Excellence: Dean's List & Vice Chancellor's List (3 consecutive terms)
  • 🚀 Production Impact: Built ML systems deployed on AWS serving real users
  • 📊 Industry Impact: Led platform revamp achieving 26% engagement improvement
  • 🔬 Conference Submissions: 2 papers under review at SIGIR 2026
  • 🏅 MSc Dissertation: Custom LLM fine-tuning for conversational search

📚 Publications & Research

  1. Rayhan, M., Ullah, M. Z. (2026). Early Identification of Ambiguous Search Queries Using Adversarially Robust Large Language Models. Submitted to SIGIR 2026 (Long Track). [Under Review]

  2. Rayhan, M., Ullah, M. Z. (2026). Adversarial Supervised Fine-Tuning for Robust Ambiguity Detection in Queries for Conversational Search. Submitted to SIGIR 2026 (Short Track). [Under Review]

  3. Rayhan, M., Alam, M. G. R., Dewan, M. A. A., & Ahmed, M. H. U. (2022). Appraisal of high stake examinations during SARS-CoV-2 emergency with responsible and transparent AI: Evidence of fair and detrimental assessment. Computers and Education: Artificial Intelligence. [Q1, Cited: 4] 📄 Paper


💡 What I'm Learning

graph LR
    A[Current Focus] --> B[RAG Architectures]
    A --> C[Vector Databases]
    A --> D[LLM Production Deployment]
    A --> E[Prompt Engineering]
    B --> F[FAISS/Pinecone]
    C --> F
    D --> G[TensorRT/ONNX]
    E --> H[Chain-of-Thought]
Loading
  • 🔍 RAG Systems: Building production-grade retrieval-augmented generation pipelines
  • Model Optimization: TensorRT, ONNX for efficient inference
  • 🗄️ Vector Databases: FAISS, Pinecone, Weaviate for semantic search
  • 🎯 Advanced Prompting: Chain-of-thought, few-shot learning, prompt optimization
  • 🏗️ LLM Agents: Building autonomous systems with LangChain & LlamaIndex

🤝 Let's Connect

LinkedIn Email Google Scholar ResearchGate


📌 Pinned Repositories

⭐ Featured Projects Coming Soon:

  • 🔍 RAG-Based Document QA System
  • 🤖 LLM Fine-tuning Pipeline (PyTorch)
  • 📊 Conversational Search Dataset & Benchmarks
  • 🎯 GPU-Accelerated NLP Toolkit

Profile views

Built with 💙 by Rayhan | Powered by curiosity and countless cups of tea ☕

@importrayhan's activity is private