Demid Valiullin / Graz, Austria

Software engineering with a mathematician’s eye for detail

I’m a mathematics student at TU Graz working on AI-assisted software projects in Python and TypeScript. I define the problem, make product decisions, and review and test the result.

I’m exploring student software development and test automation roles across Austria: 8 to 12 hours per week during the semester, or a full-time summer internship in 2027. I’m open to on-site or hybrid work and remote teams that can hire from Austria.

Graz, Austria / TU Graz

Start with the problem. Build what helps. Test what matters.

I start with the person who needs the result and the decision they need to make. Then I build the simplest version that solves the real problem.

  1. Understand the problem

    I ask what the result is for, what information it needs, and what could go wrong.

  2. Build a working path

    I work with AI coding tools on Python and TypeScript prototypes, with explicit inputs, decisions, and failure states.

  3. Test and improve

    I automate repeatable checks, review the result by hand, and use feedback to sharpen the next version.

Selected projects

Projects I can show and explain.

These independent prototypes were built with substantial AI assistance. I chose the problems, shaped the workflows, reviewed and tested the implementations, and decided what was ready to publish.

Decision support and QA automation

TenderVerdict Next Gen

Result
128 Python tests and 20 native contract checks

TenderVerdict helps a supplier review public tenders. It compares notice metadata with a supplier profile, explains each match, and keeps the first decision local.

What I did
I decided how the tool should qualify a notice, set the release scope, and checked the Python and Swift builds before submitting the project.
Built with
Python, Swift, SwiftUI, RevenueCat Test Store
Status
Hackathon prototype; not released on the App Store or used in production
Public materials
Devpost entry, competition source branch, and demo video.

Quality control for learning content

LessonProof

Result
42 automated tests

LessonProof is a prototype for safer AI edits to teaching materials: AI suggests a change, a person approves it, and automated checks stop a bad release.

What I did
I designed the approval flow and release checks, then reviewed the implementation against 42 automated tests.
Built with
TypeScript, React, Node.js
Status
Public prototype
Public materials
Runnable prototype and public source code.

Source-linked preparation checks

Neja Ready

Result
79/79 verification checks

Neja Ready turns past yacht-charter instructions into a practical checklist. It links each step to its source, calls out missing information, and runs a preflight check.

What I did
I defined what the checklist had to cover and how it should flag gaps, then verified the public prototype against all 79 checks.
Built with
TypeScript, Node.js
Status
Public prototype
Public materials
Runnable prototype and public source code.

Mathematics education

Demid.Math

Result
Public website, channel, and teaching videos

Demid.Math is where I teach mathematics in Graz. I create structured explanations and exam-preparation materials with German terminology and full worked solutions.

What I did
I choose the topics, prepare and record the material, publish it, and keep the project running.
Built with
Curriculum design, video, bilingual communication
Status
Active teaching project
Public materials
Website, Telegram channel, and mathematics videos.

Early in my software career. Serious about the work.

Demid Valiullin

I live in Graz and study mathematics at TU Graz. I’m looking for my first software engineering role, and I use public projects to show how I think, build, and test.

In those projects I’ve used Python, TypeScript, REST APIs, SwiftUI, and automated validation. When a problem calls for an unfamiliar tool, I’m comfortable learning it.

Running Demid.Math has also taught me to take responsibility for a public service, explain hard ideas clearly, and improve through direct feedback.

What teammates can expect

  1. Ask questions before a wrong assumption turns into rework.

  2. Share work early enough for feedback to be useful.

  3. Test the risky parts and explain the trade-offs.

Looking for a working student who learns fast and cares about the details?

During the semester, I can work 8 to 12 hours per week. I’m also interested in a full-time summer internship in 2027, across Austria. See my code, download the CV, or message me on LinkedIn.