AIworthiness

About me

Göksu Anıl Kızıldoğan

Principal Data Scientist · PhD Researcher, Aviation Technologies

LinkedIn profile Email

I have spent more than ten years turning data into working systems — first as a performance and calibration engineer on aerospace and piston engines at TEI, then leading data science and IoT analytics teams that ship machine learning into production.

Today my focus is large language models: retrieval-augmented generation, fine-tuning, MCP-based agents and on-premise LLMs for sensitive enterprise data. In my PhD at Eskişehir Osmangazi University I apply these techniques to aviation documents and ask what it takes for AI to be airworthy: grounded, traceable and ready for certification.

Experience

  1. Senior Data Scientist & IoT Analytics Practice Lead

    Analythinx | The Data Science CompanyJanuary 2023 – Present

    • Led the company's IoT analytics initiatives, mentoring junior data scientists and guiding IoT-driven projects to optimize industrial and operational processes.
    • Developed end-to-end Large Language Model (LLM) solutions, specializing in Retrieval-Augmented Generation (RAG) and fine-tuning for enterprise AI applications. Engineered custom AI agents using Model Context Protocol (MCP), integrating structured and unstructured data pipelines that turn raw data into embeddings for QnA systems and chatbots.
    • Designed and deployed on-premise, company-specific LLMs for sensitive data, while leveraging OpenAI APIs and models such as DeepSeek, LLaMA, Gemini and Claude when required.
    • Executed predictive analytics and ML/DL projects across production sites, manufacturing plants, human resources and CRM analytics — time series forecasting, anomaly detection and deep learning architectures.
    • Architected and deployed scalable machine learning pipelines across cloud and edge computing, applying MLOps best practices for model lifecycle management, version control and real-time monitoring.
    • Collaborated with cross-functional teams to translate business problems into data-driven solutions.

  2. Lead Data Scientist

    Analythinx | The Data Science CompanyOctober 2021 – January 2023

    • Led end-to-end data science projects, building and deploying scalable models with Python, SQL and NoSQL on robust data pipelines and ETL workflows.
    • Developed regression, classification, time series and unsupervised models using XGBoost, LightGBM, Random Forests, Transformer-based models and RNNs, optimized through Bayesian hyperparameter tuning and feature engineering.
    • Built deep learning applications with PyTorch and TensorFlow, bringing the latest research into production-ready models.
    • Architected MLOps pipelines with Airflow for orchestration, MLflow for experiment tracking and Google Cloud Platform for deployment, optimizing for latency, efficiency and cost.
    • Conducted statistical analysis, Bayesian inference and hypothesis testing; applied feature selection and dimensionality reduction to improve interpretability.
    • Mentored a team of junior data scientists and engineers through code reviews, model debugging sessions and ML workshops; led onboarding and training.

  3. Lead Performance & Data Science Engineer

    TEI Inc. — General Electric Assoc.July 2017 – October 2021

    • Developed simulation pipelines integrating 1D (GT-Power) and 3D (ANSYS Forte) modeling tools with MATLAB & Simulink for automated engine performance analyses and optimization.
    • Built data-driven calibration processes for piston engines (four-stroke diesel and two-stroke gasoline) using ATI Vision and ETAS INCA, with Python tooling for calibration and data processing.
    • Applied statistical methods — regression analysis and probabilistic models such as 3-sigma analysis — to improve engine calibration accuracy.
    • Automated extraction, transformation and analysis of calibration and test data with Python, MATLAB and DIAdem; produced reports with AVL Concerto and ETAS MDA.

Education

  • PhD, Aviation Technologies

    2024 – Present

    Eskişehir Osmangazi University

    Thesis: Python application of RAG and fine-tuning techniques on aviation documents

  • MSc, Mechanical Engineering

    2020

    Eskişehir Osmangazi University

    Thesis: Control algorithm modeling of power split series-parallel hybrid engine system · Ranked 1st of 22

  • BSc, Mechanical Engineering

    2017

    Middle East Technical University (METU)

    Implemented a custom 3D slicer in Python for 3D printer applications