tanzir.

Loading portfolio...

hi there👋, I'm

|

25, he/him

|

about me.

I'm an AI Engineer who builds smart, efficient solutions to complex problems. I specialize in machine learning, natural language processing, and automation—developing predictive tools, streamlining data systems, and connecting platforms like vector databases, LangGraph, and n8n. Whether you need to integrate custom automation into your database or fine-tune an open-source model for a specific task, I can help make it happen.

  • BanglaNewsClassifier: A machine learning approach for news classification in Bangla Newspapers Published in Q1 Journal

    Collaborators: Ar-Rafi Islam, Md Humaion Kabir Mehedi, Annajiat Alim Rasel, M. Abdullah-AL-Wadud, Jia Uddin

    View Paper

    Developed a high-accuracy Bangla news classification approach using hybrid models, achieving 94% accuracy with a BiLSTM-SVM stacking classifier on 118,404 articles, advancing low-resource language processing. Published in PLOS ONE, 2025.

  • Advancing legal accessibility in Bangladesh through AI-powered assistance and natural language interfaces Published - Undergraduate Thesis

    Collaborators: Md. Shakil Anawar, Ar-Rafi Islam, Md. Adnan Karim

    View Paper

    Created a legal question-answering system combining transformer-based retrieval, rule-based filtering, and human-in-the-loop feedback. It enabled structured access to Bangladeshi legal content via LLM-based answers and legal citation mapping.

  • joinventureai.com

    Jr. AI Developer Dhaka, Bangladesh

    Jul, 2025 - Present

    • Developed ML and neural network models with TensorFlow and PyTorch; exported via ONNX for cross-platform use.
    • Fine-tuned language models using LoRA for financial data analysis and trend prediction.
    • Fine-tuned language models using Unsloth for understanding and trend prediction.
    • Automated AI workflows and inference pipelines using n8n.
    • Created FastAPI endpoints integrated with n8n in a hybrid setup for AI automation tasks.
    • Python
    • TensorFlow
    • PyTorch
    • ONNX
    • LoRA
    • Unsloth
    • n8n
    • FastAPI
    • Git & Github

  • timerni.com

    AI & ML Research Intern Dhaka, Bangladesh

    Jan, 2025 - Jun, 2025

    • Contributed to the research and development of novel machine learning models, leading to a 10% improvement in predictive accuracy for internal datasets.
    • Fine-tuned LLMs (Mistral-7B, LLAMA 3.1) using LoRA/QLoRA, PEFT, and Hugging Face.
    • Developed RAG pipelines for NLP-based QA.
    • Built preprocessing pipelines (SentenceTransformers) for Bangla-English bilingual text.
    • Automated PDF parsing and OCR-based data extraction.
    • Authored and maintained comprehensive documentation for experimental setups, data pipelines, and model architectures to ensure reproducibility and knowledge sharing within the team.
    • Collaborated with senior researchers to analyze and interpret complex datasets, presenting key findings and insights during weekly team meetings.
    • Python
    • Mistral-7B
    • LLAMA 3.1
    • LoRA
    • QLoRA
    • PEFT
    • Hugging Face
    • RAG
    • NLP
    • SentenceTransformers
    • OCR

  • BanglaRAG - Retrieval-Augmented Generation

    RAG pipeline combining FAISS vector search with Bangla document embeddings for accurate information retrieval.

    • FAISS-based vector search implementation
    • Bangla document embedding generation
    • Grounded answer generation for Bangla queries
    • Enhanced accuracy for low-resource language
    • Retrieval-augmented LLM integration
    PythonFAISSRAGLLMsSentenceTransformersBangla NLP

  • Agent Whisper App

    live preview

    Real estate agent platform with agentic matchmaking using n8n, Softr, and Airtable for intelligent listing management.

    • Softr frontend with Airtable backend integration
    • Agentic matchmaking using n8n automation
    • Rule-based filtering for buyer-listing matching
    • Video generation from static images with FFmpeg
    • Real-time updates and user interactions
    • Hosted on Google Cloud Platform
    n8nSoftrAirtableFFmpegGCPAutomation

  • Apply Job Agent

    live preview

    ML-driven candidate screening platform using resume parsing, text embeddings, and vector databases for intelligent job matching.

    • Resume parsing and text embedding generation
    • PostgreSQL with SQLAlchemy for data management
    • Vector database integration for semantic search
    • Automated reporting workflows
    • Evaluation pipeline for continuous improvement
    • Personalized job matching algorithm
    PythonPostgreSQLSQLAlchemyVector DatabasesText EmbeddingsML

  • BRAC University

    Bachelor of Science in Computer Science

    2020 - 2024

    Dhaka, Bangladesh

    3.40 CGPA

    Achievements:

    • ★Dean's List (2020)
    • ★Academic Merit Scholarship (2021)

    Extra Curriculars:

    • •Senior Executive at BRAC University Computer Club
    • •Peer Reviewer: PLOS ONE Journal (Natural Language Processing papers)

  • M.E.H Arif College

    Higher Secondary Certificate (HSC) - Science Stream

    2016 - 2018

    Gazipur, Bangladesh

    4.58 GPA

  • Mouchak Scout High School

    Secondary School Certificate (SSC)

    2014 - 2016

    Gazipur, Bangladesh

    5.00 GPA

  • Machine Learning with Tree-Based Models in Python 2024Certification

    Credential

    Mastered decision trees and ensemble methods for classification and regression using scikit-learn. Learned bagging, random forests, AdaBoost, and Gradient Boosting. Gained expertise in hyperparameter tuning and addressing the bias-variance tradeoff in tree-based models.

  • Cluster Analysis in Python 2024Certification

    Credential

    Gained proficiency in unsupervised learning through clustering algorithms using SciPy. Covered hierarchical clustering, k-means clustering, data preprocessing, and visualization techniques. Applied clustering to real-world problems including image color analysis and news article grouping.

  • Dean's List - BRAC University 2020Award

    Recognized on the Dean's List at BRAC University for outstanding academic performance in 2020.

  • Academic Merit Scholarship - BRAC University 2021Award

    Recipient of Academic Merit Scholarship at BRAC University for exceptional academic achievement in 2021.

  • Languages:

    Bengali (Native), English (Fluent), Hindi (Conversational)

  • Programming Languages:

    Python, JavaScript, C++, SQL

  • Machine Learning Frameworks:

    TensorFlow, PyTorch, Scikit-learn, LangChain, LangGraph, FastAPI

  • AI/ML Models:

    LLM Fine-tuning (LoRA, QLoRA, PEFT, Unsloth), Neural Networks, RAG Pipelines, Transformers

  • Databases:

    MySQL, PostgreSQL, Qdrant, FAISS, Airtable

  • Web Technologies:

    FastAPI, Django

  • Deployment & Scaling:

    Docker, GCP (Google Cloud Platform), AWS

  • Tools & Libraries:

    Pandas, Scikit-learn, SQLAlchemy, FAISS, Git, Linux, Jupyter, Hugging Face

  • Natural Language Processing:

    OCR, Document Classification, Text Parsing, SentenceTransformers, Bilingual Text Processing

  • Additional Technologies:

    n8n, Softr, ONNX, FFmpeg

Let's work together.

I'm always interested in new opportunities and exciting projects. Whether you have a project or a research idea in mind or just want to chat about tech, I'd love to hear from you.

© 2025 Tanzir Hossain. All rights reserved.