Curriculum vitae

Cenk
Erdoğan.

I build AI tools for enterprise knowledge management and research explainable AI for microscopy.

What I’m doing now

AI Powered Operations

I’m building AI infrastructure and an internal knowledge base to support enterprise knowledge management, making it easier for teams to find and use information across internal sources.

Master’s thesis · Explainable AI for bioimage analysis

I’m developing an explainable workflow to profile microscopy images of Parkinson’s disease models in 2D neural stem cell cultures and 3D brain organoids. I use cell segmentation, feature extraction, machine learning and SHAP to make model decisions easier to interpret.

I compare interpretable morphological features with learned image representations to understand what each captures in the microscopy data.

Explore my work

From MATSE to cloud-ready software

My dual degree combined applied mathematics, computer science and hands-on software development. For my thesis, I broke a nightly supply-chain batch process into modular containers and orchestrated them with Kubernetes.

Read the education entry

How the nightly batch changed

Cloud Engineering

At INFORM I worked on CI/CD and Kubernetes deployments on Azure. For my own portfolio, I use a smaller setup. The diagram shows how a Git push reaches the website you’re reading.

Read the experience entry

From Git push to website

    Applied AI for internal knowledge

    I built a LangGraph multi-agent system that brings together source code, Jira and Confluence. I also designed evaluations to check whether it retrieves the right evidence and uses it in its answers.

    Read the experience entry

    How a question becomes an answer

    Illustrative example using fictional documents.

    “Where is the deployment check documented?”

    Research & Robotics

    In an autonomous systems course, my team simulated a differential-drive robot and used an Extended Kalman Filter to estimate its position from noisy sensors. Try the simplified browser demo to see why the filter needs to account for measurement noise.

    View the robotics project

    Robot localization

    A simplified Kalman filter demonstration.

    • True path
    • Noisy measurements
    • Estimated path

    Experience

    AI-Supported Enterprise Knowledge Management

    • Developed a LangGraph multi-agent system for Jira ticket analysis, codebase Q&A and documentation retrieval.
    • Combined source code, Jira and Confluence in a retrieval-augmented generation pipeline using vector search.
    • Analyzed developer workflows and evaluated retrieval and answer quality using reference questions and stakeholder feedback.

    Working Student · Software Engineering and DevOps

    • Automated builds, tests and Helm deployments to Azure Kubernetes Service with GitHub Actions and built Docker images for internal services.
    • Restored Grafana dashboards and Loki/Mimir monitoring pipelines. Automated cleanup of unused Azure Container Registry images.
    • Helped provision Jupyter infrastructure for reproducible Python workflows and wrote infrastructure tests using CDK for Terraform.

    Mathematical-Technical Software Developer Trainee

    Developed and tested software using C++, T-SQL and Visual Studio. Supported GitHub Actions and TeamCity CI/CD workflows and helped onboard new trainees and interns.

    Internship

    School Internship

    Education

    M.Sc. Artificial Intelligence

    Course projects span machine learning, autonomous systems, multi-agent systems and knowledge graphs.

    B.Sc. Applied Mathematics and Computer Science

    Dual-study MATSE program. For my thesis, I designed a cloud-ready operating model for the ADD*ONE nightly batch process. The thesis and oral defense each received a grade of 1.0 with distinction.

    Certified Mathematical-Technical Software Developer

    IHK qualification completed alongside my degree. My final project was a Java solver for a 3D wooden puzzle.

    High School Diploma

    Final grade of 1.1. Valedictorian.

    Research and awards

    Citation Farming on ResearchGate

    I presented our research on identifying coordinated citation boosting from citation-network structure. Our poster was selected as a candidate for the Best Poster Award.

    Maastricht Borderless · Maastricht Track Winner

    My team built an AI-assisted platform to help students and young professionals navigate cross-border life in the Netherlands, Germany and Belgium.

    View the project

    Technical skills

    Programming
    Python, Java, TypeScript, JavaScript, C++, C#, SQL, Bash, PowerShell
    AI and LLMs
    LangGraph, LangChain, multi-agent systems, RAG, prompt engineering, LLM evaluation, vector search, embeddings
    ML and computer vision
    scikit-learn, NumPy, scikit-image, Cellpose, StarDist, image segmentation, feature extraction, SHAP
    Data and APIs
    Knowledge graphs, RDF, OWL, SPARQL, RDFLib, REST APIs, JSON, YAML
    Cloud and DevOps
    Azure, AKS, Docker, Kubernetes, Helm, Terraform, CDK for Terraform, GitHub Actions, Azure DevOps
    Infrastructure
    Grafana, Loki, Mimir, JupyterHub, Git, Linux, software architecture, test automation

    Languages

    German
    Native
    English
    Full professional proficiency
    Turkish
    Professional working proficiency
    Spanish
    Elementary proficiency