Available for opportunities

Dominic
Kuncik.

Full Stack DeveloperAI Researcher

Building scalable systems at the intersection of software engineering and machine learning. PhD candidate at TU Dublin researching AI-generated content detection. Based in Dublin, Ireland.

View my workLinkedInGitHubEmail me
locationDublin, Ireland
statusPhD research + open to roles
focusAI · Full Stack · Research
01About

Developer by craft,
researcher by curiosity.

I'm a full stack developer and AI researcher based in Dublin, Ireland. My work sits at the intersection of practical software engineering and applied machine learning — I care deeply about writing systems that are fast, maintainable, and genuinely useful.

I hold a First Class Honours BSc in Computing from Technological University Dublin, and I'm currently pursuing a PhD there, researching how to detect AI-generated academic writing using personalised deep learning models.

Beyond academia, I've shipped production features at Kianda, an enterprise low-code platform, where I worked across the full stack with Ember.js, C#/.NET, and a healthy dose of debugging instinct.

2+years professional experience
PhDcandidate, TU Dublin
BScfirst class honours
currently.txt
rolePhD Researcher
institutionTU Dublin
topicAI Content Detection
sinceMay 2025
open tofull-time / contract roles
locationDublin — remote-friendly

Education

2025 →
PhD, Computer Science
Technological University Dublin
2023
BSc (Hons) Computing
Technological University Dublin
02Research
PhD Research · TU Dublin · 2025–Ongoing

Detecting AI-Generated Academic Writing

A novel system built on personalised student models, NLP, and deep learning to strengthen academic integrity in the age of large language models.

Abstract

As large language models become increasingly capable of mimicking human writing, traditional plagiarism detection methods fall short. This research proposes a fundamentally different approach: rather than detecting "AI-ness" generically, the system builds a personalised model for each student based on their unique writing history, then identifies statistically significant deviations in grammar, vocabulary, and stylistic structure.

Personalised Baselines

Each student's prior work trains an individual writing profile, making the system robust to general style variation between students.

Multi-Modal Analysis

Combines classical NLP feature extraction (NLTK) with deep learning models (PyTorch/Keras) for high-confidence classification.

Stylometric Detection

Analyses grammar patterns, vocabulary richness, sentence structure variance, and semantic coherence as fingerprint dimensions.

Mixed Techniques

Ensemble approach fusing statistical methods with neural representations to achieve robustness against prompt engineering.

Research Stack
PythonPyTorchTensorFlowKerasNLTKTransformersScikit-learnGPT EmbeddingsNLPStylometrics
03Experience

Professional History

Nov 2023 – Sep 2025
County Dublin, Ireland
Ember.jsJavaScriptC#.NETRESTful APIsMVC

Full Stack Developer

Kianda

Enterprise low-code platform serving complex business workflows across European clients.

  • Designed and developed new platform features using JavaScript, Ember.js, and C#/.NET in collaboration with cross-functional teams, enhancing scalability and usability for enterprise clients.
  • Debugged and optimised backend and frontend components to significantly improve system reliability and user satisfaction.
  • Oversaw end-to-end IT operations including asset audits and deployment of all hardware and software for staff, ensuring a secure and efficient workplace.
  • Worked directly with clients to deliver fit-for-purpose solutions, translating technical requirements into high-impact features.
Certifications
The Git & GitHub Bootcamp
Udemy · Issued September 2025
View cert →
04Projects

Things I've built

Personal ProjectLive

AlwaysRight

Settle every argument. Scientifically.

A playful web application that enables groups to log and resolve everyday debates through a structured, points-based system. Features group creation, point awarding, and a real-time leaderboard to encourage friendly competition.

PythonFastAPIReactNode.jsPostgreSQLRedisCeleryDocker
Personal ProjectBuilt

Rewardify

Customer loyalty at scale.

A high-concurrency customer loyalty engine using FastAPI and PostgreSQL to track purchases, award points, and manage redemptions. Event-driven background processing via RabbitMQ and Celery, with a modern React dashboard containerised in Docker.

PythonFastAPIReactPostgreSQLRabbitMQCeleryDocker
Personal Project / Open SourcePublished

Hevy API Wrapper

Typed Python client for the Hevy API.

A fully-typed, comprehensive Python client for the Hevy fitness tracking API, supporting both synchronous and asynchronous usage patterns. Designed with clean interfaces and full type annotations for developer ergonomics.

PythonREST APIsAsyncIOType Hints
University ProjectGrade A

AI-Powered ANPR Data Collection

Gamified crowdsourcing with computer vision.

University capstone project (Grade A). Designed a system to gamify data collection using object detection, Automatic Number Plate Recognition (ANPR), and web interfaces to automate and improve data quality.

PythonTensorFlowOpenCVFlaskMySQLYOLO
University ProjectGrade A

NFC Smart Attendance

Replace paper registers with a tap.

NFC-based mobile attendance system developed during COVID-19 to replace paper registers. Android application handles NFC registration and automated attendance logging, with a MySQL backend.

JavaKotlinAndroid StudioNFCMySQL
05Skills

Technical Toolkit

Languages & Runtimes

PythonJavaScriptTypeScriptJavaCC#C++SQL

Frontend

ReactNext.jsEmber.jsHTML5CSS3

Backend & APIs

FastAPIFlaskNode.js.NETRESTful APIsMicroservicesMVC

Databases

PostgreSQLMySQLSQLAlchemyAlembicRedis

AI & Machine Learning

PyTorchTensorFlowKerasNLTKOpenCVYOLOResNetVGGLarge Language ModelsNLPComputer VisionModel Training & Evaluation

Python Ecosystem

NumPyPandasPydanticJinjaCeleryRabbitMQ

DevOps & Infrastructure

DockerGitGitHubCI/CD
06Contact

Let's work
together.

I'm open to full-time roles, contract work, and research collaborations. Whether you're building something ambitious or looking for a developer who can bridge the gap between applied AI and production software — I'd love to hear from you.

Response time is usually within 24 hours.

Email
itsdomhd@gmail.com
LinkedIn
linkedin.com/in/kuncik
GitHub
github.com/dkuncik
Available for new opportunities