A young professional with an MSc degree in Informatics from TUM specializing in Machine Learning and Computer Vision, with 3+ years of work experience. Passionate about deploying cutting-edge ML solutions to solve real-world challenges and bridge the gap between research and practical applications.

Work Experience
  • •Designed, developed, and deployed secure HTTP servers and APIs using Node.js on AWS.
    • 2024/10 – presentMunich, Germany
  • •Experimented with deploying local LLM by using llama and fastapi.
  • •Optimized data models to enhance performance and scalability in Tableau and Power BI.
  • •Consulted directly with customers, delivering tailored dashboards and deploying data-driven solutions.

    • Student Assistant @ GRIS-MEC Lab

    Technische Universität Darmstadt, "One of the best German universities"
  • •Experimented with SGDiff and conducted baseline experiments for surgical scene generation with diffusion models and scene graph conditioning.
    • 2024/04 – 2024/10Remote, Germany

    Machine Learning Thesis Student @ Research & Technology

    ZEISS, "World leading medical technology company"
  • •Developed a method, SANGRIA, an annotation-efficient method for surgical video workflow analysis.
    • 2023/07 – 2024/02Munich, Germany
  • •Proposed unsupervised video object segmentation method on surgery videos.
  • •Created dynamic scene graph representations for surgical videos.

    • Machine Learning Working Student @ Research & Technology

    ZEISS, "World leading medical technology company"
  • •Researched unsupervised object discovery with foundation models on surgical scenes.
    • 2022/09 – 2023/07Munich, Germany
  • •Developed weakly supervised instance segmentation model on surgery videos.
  • •Improved surgical tool segmentation performance by ~2%.

    • Research Assistant @ KUIS AI Lab

    Koc University, "Top research university of Turkey"
  • •Researched applications of object detection, domain adaptation, and self-supervised learning on comic scenes.
    • 2020/09 – 2021/10Istanbul, Turkey
  • •Experimented with Swin Transformers, DINO, MoCo v2, and Jigsaw Puzzle task in self-supervision research.
  • •Explored image generation with Generative Adversarial Networks (GANs) in domain adaptation research.
  • •Leveraged MMdetection and Detectron2 frameworks for object detection pipelines.
  • •Developed a self-labeling tool on top of LabelImg to ease the annotation process.

    • NLP Working Student @ AI Team

    Akbank, "Most valuable banking brand of Turkey in 2018"
  • •Developed data pipelines for efficient data crawling, text preprocessing, and tokenization, enhancing data processing speed and accuracy.
    • 2019/10 – 2020/06Istanbul, Turkey
  • •Improved performance of spellcheck and autocomplete services.
  • •Constructed knowledge graph by using Neo4j enabling data relationships and insights.
    • Education

    Master of Science, Informatics

    Technical University of Munich

    1.7/​5.0 (German GPA).

    2021/04 – 2025Munich, Germany

    Master Thesis Topic: Annotation Efficient Surgical Video Analysis with Graph Machine Learning (Grade 1.0)

    Bachelor, Computer Science

    Sabancı University

    3.62/​4.00 (USA GPA)

    2015/09 – 2020/06Istanbul, Turkey
    Languages
    English

    C1

    German

    B1

    Caghan Köksal
    1 / 3
    Publications

    Semantic Scene Editing for Cholecystectomy Surgery (COLAS @MICCAI25)

    Çağhan Köksal, Yousef Yeganeh, Azade Farshad, Nassir Navab
  • •We propose an image editing framework tailored for surgical scene manipulation.
    • 2025
  • •We show that our model enables fine-grained editing (translation, rotation, tool replacement, and tool removal) based on triplet prompts and binary tool masks.

    • SurGrID: Controllable Surgical Simulation via Scene Graph to Image Diffusion (IPCAI 25)⁠

    Yannik Frisch, Ssharvien Kumar Sivakumar, Çağhan Köksal, Elsa Böhm, Felix Wagner, Adrian Gericke, Ghazal Ghazaei, Anirban Mukhopadhyay
  • •Surgical scene generation with diffusion models with scene graph conditioning.
    • 2025
  • •Received the Joint Winner - Snke⁠OS Machine Learning in CAI Award
  • •End2end surgical workflow prediction solution with dynamic scene graph representations based on spectral clustering.
    • 2024
  • •The paper was based on my thesis work at TUM & Carl Zeiss AG and was published in the GRAIL Workshop @MICCAI 24. We got the best paper award.

    • SURGIVID: Annotation-Efficient Surgical Video Object Discovery (Arxiv 24)

    Çağhan Köksal, Ghazal Ghazaei, and Nassir Navab
  • •Investigated semi-supervised learning for surgical scene segmentation, leveraging object discovery and weak supervision to reduce annotation dependency.
    • 2024

    SU-NLP at TREC NEWS 2020 (TREC 2020)⁠

    Ali Eren Ak, Çağhan Köksal, Kenan Fayoumi, and Reyyan Yeniterzi

    BERT Summarization, Universal Sentence Encoder, and BERT Finetuning approaches are examined in Background Linking tasks and Wikification.

    2020

    My contributions are :

  • •Setting up the baseline result with ElasticSearch.
  • •Experimenting with Universal Sentence Encoder embeddings in the Background Linking task.

    • Projects

    Visual Question Answering on Medical Data

  • •Worked on large-scale multimodal datasets, MIMIC-CXR and ROCO.
  • •Used SOTA multimodal Flamingo architecture.
  • •My contributions:
  • •Implemented data pipelines in VQARAD, ROCO, MIMIC-CXR, and ImageCLEF datasets.
  • •Experimented with Vision Transformers and EfficientNet.
  • •Used in-domain language models such as PubMedBERT, GPT2 for feature extraction and text generation.
  • •Created training and evaluation pipelines of VQA and text generation tasks.
  • •Modeled VQA as a multi-task learning problem and improved VQA performance by %2.

    • Graph Neural Networks on Abdominal Data Meshes

    Explore Graph Neural Networks on abdominal data meshes.

  • •Implemented and experimented with GCN, GraphSage, and FeastNet models.
  • •Created data, training, and evaluation pipelines.
  • •Worked on gender prediction, age, BMI, height, and weight regression tasks.

    • Self-Supervised-Learning on COMICS

    Explored vision based self-supervision methods to assess their capabilities on highly stylistic comic domains. The following methods are used in my experiments:

  • •DINO (Vision Transformer and Swin Transformer), MoCo v2, Jigsaw task, attention map visualization, and input gradient-based method are used to understand the learned features.

    • Self Labeling Tool based on LabelImg

  • •Developed a self-labeling tool for object detection.
  • •Powered by FasterRCNN of MMDetection framework
  • •Automatically generate bounding boxes and let users update them.

    • Developing a Vehicle Price Prediction Model Using Existing Features and Available Text

    Bachelor Thesis
  • •Secondhand car price prediction system that uses the text of the car ads and structured data such as mileage, brand.
  • •The dataset is created by crawling the website of one of Turkey's biggest second-hand car sellers.
  • •Tech: Selenium, UiPath, BS4
  • •Word2Vec embeddings are used to examine the relationship between the text's brands, features, and correct typos.
  • •Doc2Vec and CNN-based approaches are used to vectorize the text.
    • Caghan Köksal
    2 / 3
  • •Linear Regression, Decision Tree, and Random Forest methods are used for the final regression task.
    • Caghan Köksal
    3 / 3