Michel Yildirim

Junior Software Engineer · Morch Solutions

Aarhus, Denmark

• Work across a portfolio of clients, from small businesses to enterprise, often switching between codebases, stacks and domains within the same day.
• Field-service platform for cable and broadband providers: technician booking and scheduling, job tracking, equipment orders and Wi-Fi labels.
• The full IT system for a rehearsal-space business: bookings for bands, musicians and teachers, check-in, door codes, room maintenance and their expansion into podcast studios.
• Smaller end-to-end web apps, and well-scoped technical tasks such as API integrations for enterprise clients.
• Agentic development as the default way of working: coding agents do much of the implementation, and my time goes into scoping, review and verification.
• CI/CD, deployment and DevOps: automated build, test and release pipelines in GitLab, and Docker.

Research Assistant · CLAN, Aarhus University

Aarhus, Denmark

• No Hands: research on programming by voice, for developers with motor impairments.
• Set up and ran a user study on how well current AI tools support voice programming.
• Building a prototype VS Code extension that turns speech into code edits and editor commands using LLM tool use.

Teaching Assistant · Aarhus University

Aarhus, Denmark

• Taught 25+ students per semester in object-oriented programming in Java.
• Wrote an automated runner that tests student submissions against test suites to support grading.
• Taught introductory object-oriented modelling (UML) and JavaScript in a digital fundamentals course.

Danish Championship in AI · solo entry

• Tumour segmentation in medical images (computer vision).
• Reinforcement learning for a simulated car driving on sensor input.
• A retrieval-augmented LLM pipeline for healthcare questions, tuned for accuracy and speed.

Nitor Energy Quantitative Trading Case Competition

• Forecasting models on two years of hourly European power-grid data. Placed 12th of 45 teams.

Wikinsert · bachelor project

• An entity-aware insertion tool for Wikipedia editors, built with CLAN at Aarhus University.

Incommodities Case Crunch

• A LightGBM pipeline forecasting wind-energy yield from spatio-temporal weather data.

AUHack

• Simulated CPU/GPU scheduling with dynamic priority queues. Honorary mention among 30+ teams.

MSc Computer Science, specialising in AI · Aarhus University

Aarhus, Denmark

Exchange semester, School of Computing · National University of Singapore

Singapore

• Reinforcement learning and sequential decision making; smart systems and AI governance.

BSc Computer Science · Aarhus University

Aarhus, Denmark

• Machine learning, linear algebra, compilers, distributed systems, programming languages.

International Baccalaureate · Nyborg Gymnasium

Nyborg, Denmark