Skip to content
View arpita739's full-sized avatar
:dependabot:
Building AI
:dependabot:
Building AI

Highlights

  • Pro

Block or report arpita739

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.

Please don't include any personal information such as legal names or email addresses. Maximum 100 characters, markdown supported. This note will be visible to only you.
Report abuse

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

Report abuse
arpita739/README.md

πŸ‘‹ Hi, I'm Arpita Halder

AI Developer | Data Scientist | Researcher | Open-Source Contributor

Welcome to my GitHub! I’m passionate about building robust AI solutions for healthcare, enterprise, and real-world applications. With experience across medical AI, document intelligence, and cloud-native ML, I specialize in turning data into impact.


πŸš€ What I Do

  • πŸ”¬ Conduct applied research in Generative AI, RAG, and Medical Imaging
  • 🧠 Build AI/ML solutions that enhance data traceability, governance, and searchability
  • πŸ› οΈ Develop end-to-end products using Python, C++, Docker, Git, Azure, and more
  • πŸ‘₯ Collaborate in Agile teams across academia, startups, and enterprise (e.g. Siemens Healthineers)

🧠 Highlighted Project

πŸ—οΈ Bit&Beam – Intelligent Building Documentation Management System

Developed during the AMOS program in partnership with BUILD.ING GmbH, Bit&Beam is an open-source platform to manage building documentation using:

  • πŸ“„ AI-powered document classification
  • πŸ” Smart metadata extraction and NLP search
  • πŸ›‘οΈ Secure multi-tenant access and OCR capabilities

πŸ’‘ Built in a cross-functional Scrum team with industrial stakeholders to deliver a real-world solution.


πŸ“ Publications

  1. COVID-19 Detection from Lung CT-Scan Images using Transfer Learning
    Machine Learning: Science and Technology (IOP Science)
    Achieved 97% accuracy and 0.99 AUC using DenseNet on lung CT scans.
    πŸ”— GitHub

  2. Real-time Vernacular Sign Language Recognition
    IJRPR Journal
    Built a real-time system using MediaPipe and ML, achieving up to 99.29% accuracy.
    πŸ”— GitHub


πŸ’Ό Experience Snapshot

  • Siemens Healthineers – AI Werkstudent
    β†’ Integrated LLMs in enterprise tools for natural language querying & smart data catalogs
    β†’ Developed ML models for supply chain traceability & sustainability

  • BUILD.ING GmbH (AMOS Project) – Software Developer
    β†’ Delivered full-stack AI-powered document management system with GenAI, OCR, NLP

  • UK Erlangen – Research Assistant (Glottis & Spine Segmentation)
    β†’ Implemented segmentation pipelines for clinical image data


πŸŽ“ Education

  • πŸŽ“ M.Sc. Artificial Intelligence
    FAU Erlangen–NΓΌrnberg, Germany
    Ongoing Research: Retrieval-Augmented Generation for Medical QA (MedRAG-GPT)

  • πŸŽ“ B.Tech. Computer Science & Engineering
    MAKAUT, India
    Thesis: Signature Verification using Siamese Networks


βš’οΈ Skills

Languages & Frameworks
Python C++ SQL PyTorch TensorFlow Docker Shell OOP

Tools & Platforms
Azure Git Snowflake PowerBI PostgreSQL HPC LaTeX CI/CD

AI/ML
Generative AI LLMs RAG NLP Computer Vision Data Visualization


πŸ† Highlights

  • πŸ₯‰ 3rd Place – Healthcare Hackathon Bavaria 2024
    β†’ Built "Flora", an AI avatar to enhance patient engagement, with Siemens Healthineers

πŸ“ˆ GitHub Stats

Arpita's GitHub Stats Top Languages


πŸ† GitHub Profile Trophy



πŸ“« Get in Touch


🌍 Let’s build AI solutions that matter.

Pinned Loading

  1. Real-time-Vernacular-Sign-Language-Recognition-using-MediaPipe-and-Machine-Learning Real-time-Vernacular-Sign-Language-Recognition-using-MediaPipe-and-Machine-Learning Public

    The deaf-mute community have undeniable communication problems in their daily life. Recent developments in artificial intelligence tear down this communication barrier. The main purpose of this pap…

    Jupyter Notebook 36 8

  2. Bengali-and-Hindi-Signature-Verification-using-Convolution-Siamese-Network Bengali-and-Hindi-Signature-Verification-using-Convolution-Siamese-Network Public

    Verification of off-line signatures is one of the most challenging tasks in biometrics and document forensic science. In this thesis, we deal with Convolutional Siamese Network model which is capab…

    Jupyter Notebook 1

  3. Data-Science-Prediction-Models Data-Science-Prediction-Models Public

    In this repository you will find prediction model of different data sets taken from kaggle.com . I worked on mini Data science projects as a beginner. Hopefully this will be a stepping stone in my …

    Jupyter Notebook 1

  4. Using-GAN-fill-missing-part-of-Handwritten-Digits Using-GAN-fill-missing-part-of-Handwritten-Digits Public

    How to build a neural network to fill the missing part of a handwritten digit using GANs

    Jupyter Notebook 1

  5. amosproj/amos2025ss02-building-documentation-management-system amosproj/amos2025ss02-building-documentation-management-system Public

    A modern solution for organizing, analyzing, and retrieving building-related documents with AI-powered features

    TypeScript 3 2

  6. made-template made-template Public

    Forked from jvalue/made-template

    Template repository for the Methods of Advanced Data Engineering course at FAU

    Jupyter Notebook 1