Arpita Halder

AI Researcher

Erlangen, Bavaria, Germany2 yrs 4 mos experience

AI EnabledAI ML Practitioner

Key Highlights

Expert in AI-driven healthcare solutions.
Proficient in NLP and machine learning techniques.
Strong background in data analysis and visualization.

Stackforce AI infers this person is a Healthcare AI Developer with expertise in machine learning and data analysis.

Contact

Skills

Core Skills

Natural Language Processing (nlp)Machine LearningData ScienceData AnalysisDeep LearningComputer Vision

Other Skills

Python (Programming Language)Knowledge Graph-Based Natural Language ProcessingOntologiesLarge Language Models (LLM)Medical DatabasesTree traversal algorithmFlaskKnimeNeo4jDBeaverSnowflake CloudRAG PipelineLangChainOntology EngineeringAzure AI Studio

About

In the world of AI, change is the only constant. Early in my career, I realised that relying on a single skill or technology was never enough — the real advantage comes from learning fast, adapting quickly, and connecting the dots across disciplines. Whether it’s exploring new machine learning approaches, refining NLP models, or uncovering insights from complex datasets, I approach each challenge with both curiosity and precision. I deliver by combining rapid learning with structured problem-solving, ensuring solutions are not only innovative but also reliable, scalable, and aligned with real-world needs. For me, every project is an opportunity to grow, collaborate, and transform ideas into measurable impact.

Experience

Universitätsklinikum erlangen

Master Thesis Student

Oct 2025 – Present · 5 mos · Erlangen, Bavaria, Germany · On-site

AI-Driven Structured Reporting for Breast MRI: Leveraging LLMs for Automated Clinical Label Extraction.
It focuses on building a full pipeline to automatically extract structured, clinically coded information from unstructured German MRI reports based on the BORN template (Breast Oncology Reporting Network).
In addition, I have developed a complete AI-powered breast MRI labeling platform on top of this work — an application that assists radiologists by automatically extracting labels using LLMs and allows them to review and edit the annotations, ensuring both efficiency and clinical reliability.

Python (Programming Language)Knowledge Graph-Based Natural Language ProcessingOntologiesLarge Language Models (LLM)Medical DatabasesTree traversal algorithm+3

Siemens healthineers

2 roles

Werkstudent - Data Science and Generative AI

Promoted

Apr 2024 – Present · 1 yr 11 mos · Erlangen, Bavaria, Germany · Hybrid

Building an LLM-powered data-quality service to standardize and validate customer data for
CDP and segmentation.
Built Data engineering and ML workflows for customer analytics, covering PII pseudonymiza-
tion, funnel KPIs, and behavioral insights.
Led AI integration and generative NLP adoption in enterprise tools, enabling natural language
querying and improving data catalog usability for non-technical users.
Boosted usability of the enterprise data glossary by facilitating LLM-based enhancements for
intelligent search and navigation.
Enhanced operational efficiency by building internal applications, automating workflows,
and promoting CI/CD and Git-based development, leading to faster delivery and reduced
manual effort.
Mitigated compliance risk by up to 4% penalty avoidance through data analysis and ML so-
lutions for supply chain traceability.

Large Language Models (LLM)Deep LearningPython (Programming Language)KnimeNeo4jDBeaver+23

Werkstudent - Data Analysis

Nov 2023 – Jul 2024 · 8 mos · Erlangen, Bavaria, Germany · Hybrid

Department : SHS IT DCE MSM PDM
Conducted Thorough Data Analysis: Extracted valuable insights by employing advanced statistical techniques and data visualization tools to inform strategic decision-making processes.
Developed Applications in Microsoft Power Apps: Created robust applications to address specific internal employee needs, enhancing operational efficiency and user experience.
Applied Data Science Models: Utilized programming languages and frameworks to develop innovative business solutions, automate tasks, and streamline processes. This resulted in improved team efficiency and productivity, and included the creation of user-friendly applications to facilitate seamless workflow automation for various business-related tasks.

Python (Programming Language)Natural Language Processing (NLP)Microsoft Power BISQLData VisualizationData Analysis+4

Universitätsklinikum erlangen

Student Assistant

Sep 2023 – Mar 2024 · 6 mos · Erlangen, Bavaria, Germany · On-site

Pathologie Department
Operated various scanning devices P1000 Panoramic, P250 Flash II, Panoramic Midi II and Hamamatsu S210

Mpm micro präzision marx gmbh

Werkstudent

Apr 2023 – Oct 2023 · 6 mos · Erlangen, Bavaria, Germany · On-site

Worked on assembling different electronic hardware parts.

Fau erlangen-nürnberg

Student Research Assistant

Mar 2022 – Sep 2022 · 6 mos · Erlangen, Bavaria, Germany

ANKI Labs under Artificial Intelligence and Biomedical Engineering Department.
Worked on the Research Project: Cervical Spine Segmentation
Annotated dataset using Labelme software, implemented modified U-Net for segmentation and presented thorough literature research on Medical Image Segmentation, worked on post-processing techniques such as boundary refinement.
Assisted with Tracking Olympiad seminar with the website development and data annotation.

Universitätsklinikum erlangen

Student Research Assistant

Mar 2022 – Sep 2022 · 6 mos · Erlangen, Bavaria, Germany

Worked on Research Project : Glottis Segmentation. Implemented modified U-Net using Transfer Learning models as Encoder for feature extraction. Worked on image restoration techniques to improve the visual quality of the image.

TensorFlowImage SegmentationNumPyLabelmeImage ProcessingMicrosoft PowerPoint+6

Cska automation services pvt ltd

AI Developer Intern (Working Student)

Aug 2020 – Jan 2021 · 5 mos · India

Worked on the project "Real-time Word-Level American Sign Language Recognition from Video" using Google's Mediapipe technology and ML models.

TensorFlowImage SegmentationNumPyOpenCVPython (Programming Language)Keras+2

Indian institute of technology (banaras hindu university), varanasi

Summer Research Intern

May 2020 – Jul 2020 · 2 mos · Varanasi, Uttar Pradesh

-Worked on research project De-hazing of Images using CycleGAN. Textural quality was improved by using an optimized loss function by 20% and an observable haze-free image was obtained.

NumPyOpenCVPython (Programming Language)Machine LearningScikit-Learn