AI ENGINEER · EINDHOVEN, NL

Balázs
Csutar

I build AI systems, data pipelines and the React apps that put them in front of people_

Cum laude Data Science & AI graduate with a year of production work at two Dutch construction-expertise firms. I work in English in Dutch teams.

01 / AI

AI that shows
its work

I build AI that reads messy input, uses real tools, checks its own work, and hands over to a person when it should. Two things I build most:

SCROLL OR SWIPE FOR THE NEXT PAGE
Bitsworth{{ chatStatus }}
pixel butler · scripted, never hallucinates
{{ m.t }}
02 / DATA SCIENCE

Making data
make sense

Raw records, photos and text, turned into something people can query, read and act on.

SCROLL OR SWIPE FOR THE NEXT PAGE
{{ nGrade }}
graduation grade, out of 10
{{ nCert }}
certifications on DataCamp and other platforms
{{ nFirms }}
Dutch companies running my work in production
DATA CLEANING

Messy in, structured out

Company data arrives in every format people can think of. I ingest it, clean it and give it one structure, so employees can query figures and statistics themselves.

Python · data cleaning · PostgreSQL

DATEDAMAGECOST (€)
{{ r.a }}{{ r.b }}{{ r.c }}
Illustrative example, not real records
wall segmented crack
COMPUTER VISION

Finding the crack

An exploratory computer-vision layer that segments damage photos, so cracks can be found and categorised automatically. It’s still a prototype and not part of the MVP yet.

Computer vision · image segmentation

DASHBOARDS

Numbers people act on

A Power BI dashboard for a small webshop that tracks sales and user behaviour over time, so decisions about products and stock rest on data instead of gut feeling.

Power BI · analytics

SALES PER MONTHTREND
JFMAMJJASOND
Illustration of the dashboard. Real figures stay private.
COMMONRARE
{{ w.t }}
Relative frequency, for illustration
NLP

How hard is a word?

woordje’s Flask API analyses text with spaCy for lemmas and grammar and wordfreq for how common each word is. Frequency becomes a difficulty level, so learners see the words that matter most first.

Flask · spaCy · wordfreq · NLP

03 / REACT / NEXT.JS

Interfaces
people enjoy

Production front ends, from internal dashboards to a mobile app.

SCROLL OR SWIPE FOR THE NEXT PAGE
{{ stu.title }}{{ stu.sub }} · example data
{{ k.l }}
{{ k.v }}
{{ k.s }}
{{ stu.chartTitle }}{{ stu.legend }}
{{ stu.from }}{{ stu.to }}
{{ stu.breakTitle }}
{{ r.t }}{{ r.p }}
04 / INFRA

Under
the hood

Linux every day. Containers, Kubernetes and networking on my own home server and at school. Cloud from school projects.

SCROLL OR SWIPE FOR THE NEXT PAGE
balazs@homelab: ~
{{ tl.a }}{{ tl.b }}{{ tl.c }}
type 'help' and press enter
balazs@homelab:~$ {{ th.cmd }}
{{ ol.t }}
05 / PROJECTS

Selected
projects

Each one has its own page with the full story.

SCROLL OR SWIPE FOR THE NEXT PAGE
06 / SKILLS

The toolbox

Pick a field to see what I use for it.

SCROLL OR SWIPE FOR THE NEXT PAGE
{{ bootTxt }}
{{ tg.t }}
2023 – 2026 · Breda University of Applied Sciences
BSc Data Science & AI, cum laude
Graduation project graded 8.2/10
3
languages: Hungarian, English, Dutch (A2)
18
certifications on DataCamp and more

Get in touch

Email is quickest. I reply in English, or in simple Dutch if you prefer.

Balázs Csutar · Eindhoven, Netherlands

Csutar.
[email protected]