CV photo

Kristof Drexler

Machine Learning Engineer | MLOps

Copenhagen, Denmark

k.k.drexler@gmail.com · +45 31 49 36 71

/in/kkdrexler /dKenez dkenez.dk

About

Machine Learning Engineer working the production side of the stack — dataset generation, feature engineering, training, deployment, and monitoring. Currently building LLM pipelines at Numerator that turn large volumes of unstructured images and text into structured data products at scale, on AWS with Python, Kubernetes, Terraform, and Snowflake. Background in Computer Vision, Deep Learning, and LLMs — and the MLOps that gets them into production. The same experience is used to run a two-site personal homelab.

Languages

Hungarian

Native

English

Fluent

Danish

PD3

German, Spanish, Latin

Passionate conversationalist

Experience

Machine Learning Engineer - Numerator

2026 — Present

Copenhagen, Denmark

Numerator

Machine Learning Engineer

  • Build and operate production LLM pipelines for FMCG market analytics — owning model deployment, CI/CD, and post-deployment monitoring on AWS
  • Manage ML services on Kubernetes, IaC with Terraform, and CI/CD via GitHub Actions
  • Built a receipt-dewarping module that fits a homography from OCR word-box geometry to correct perspective ahead of LLM extraction — improving downstream extraction reliability
PythonSQLAWSSnowflakeTransformersLLMsTerraformKubernetesCI/CD
Machine Learning Engineer - Kauza ApS

2024 — 2026

Copenhagen, Denmark

Kauza ApS

acquired by Numerator, 2026

Machine Learning Engineer

  • Developed transformer-based extraction / classification for industry-leading FMCG products
  • Built data pipelines and cloud deployment infrastructure on GCP using Terraform and Docker
PythonRustGCPBigQueryTransformersTerraformDockerPostgreSQL
Lead Frontend Engineer - Metriks ApS

2022 — 2024

Copenhagen, Denmark

Metriks ApS

Lead Frontend Engineer

  • Redesigned and built a modern webapp with sleek and responsive UI/UX
  • Stepped into backend and DevOps to keep things running fast and smooth
ReactReduxNext.jsFastAPIRedisDockerTypeScriptPython
Junior Software Engineer - MaxWhere Solutions Ltd.

2021 — 2022

Budapest, Hungary

MaxWhere Solutions Ltd.

Junior Software Engineer

  • Engineered custom digital twin platforms tailored to client needs
  • Built real-time visualization tools for additive manufacturing workflows
ReactVueThree.jsElectronTypeScriptPython

Experience (cont.)

Junior Software Engineer - Profirent Kft.

2020 — 2021

Budapest, Hungary

Profirent Kft.

Junior Software Engineer

  • Automated rental management workflows with VBScript
  • Built a Python computer vision pipeline to process years of document backlogs, eliminating ~300 hours of manual work
PythonPyTorchGoogle TesseractVBScript

Education

MSc. Autonomous Systems - Technical University of Denmark

2022 — 2024

Lyngby, Denmark

Technical University of Denmark

MSc. Autonomous Systems

  • Specialized in Computer Vision and Deep Learning for Autonomous Systems
  • Thesis project: Global Localization of Earthly Images powered by Deep Learning
BSc. Mechatronics Engineering - Budapest University of Technology and Economics

2018 — 2022

Budapest, Hungary

Budapest University of Technology and Economics

BSc. Mechatronics Engineering

  • Specialized in Cyberphysical Systems and Robotics
  • Thesis project: Implementing Additive Manufacturing Processes in Digital Twin Systems

Projects and Achievements

2025 — Present

Budapest / Copenhagen

edda — Two-site self-hosted infrastructure

Personal project

  • Two-site homelab (Budapest / Copenhagen) and a cloud edge node, meshed over Tailscale; both sites behind CGNAT with a single public entry point and split-horizon DNS
  • Fully reproducible from code: Terraform for guest lifecycle, Ansible for OS and service config; any node rebuilds from a minimal OS image with one playbook run
TerraformAnsibleProxmoxZFSk3sDockerTailscalePrometheus
MSc. thesis project - LOCUS Image Geolocation Webapp

2024

Copenhagen, Denmark

LOCUS Image Geolocation Webapp

MSc. thesis project

  • Full stack Deep Learning project with webapp deployed on the cloud
  • Implemented a novel method for global image localization using deep learning
  • Developed using cutting-edge technologies in MLOps, Webdev, and Deep Learning
SvelteKitTailwindCSSTypeScriptPyTorchPolarsPostgreSQLDocker
Machine Learning and Data Mining | MLOps | Introduction to Programming - Teaching Assistant at Technical University of Denmark

2023

Lyngby, Denmark

Teaching Assistant at Technical University of Denmark

Machine Learning and Data Mining | MLOps | Introduction to Programming

  • Assisted students in understanding complex topics in Machine Learning and Data Mining
  • Communicated complex ideas clearly across skill and language barriers