Sadeq RezaiConfiguration / Quality / AI Let’s talk

Sadeq Rezai Budapest, Hungary

Complexity,made useful.

I configure, connect and validate AI-enabled systems—turning unclear requirements into workflows people can actually use.

01 / From fragments to a working system

Scroll to explore · original paper-cut artwork

Different disciplines.
One connected practice.

01
ConfigurationGive it structure.
02
ValidationMake it testable.
03
AI-enabled deliveryMake it useful.

01 / Selected work

Built to be
put to the test.

Browser checks, agent workflows and data products. Different systems, with the implementation and its limits open to inspection.

Peoples Clinic monitoring dashboard showing workflow execution history

Actual project screenshot · historical results, not live status

01.1 / Quality engineering

Reliability you
can inspect.

The Peoples Clinic QA Suite brings browser journeys, API checks and AI-feature validation into a scheduled monitoring workflow.

10 minDocumented smoke-test schedule
3 suitesSmoke, regression, AI & parser
Privacy-safe illustration of the Seneschal workspace with agent and approval panels

Public product illustration · beta

Seneschal

A local-first workspace for directing AI agents, with explicit permissions, visible handoffs and review.

01.2 / AI workflow tools · Windows & WSL

Exchange Universe R Shiny application with currency charts and comparisons

Actual application screenshot

Exchange Universe

More than 170 currencies, explored through an R Shiny application built around comparison and context.

01.3 / Data products · R / Shiny / APIs

The source matters as much as the surface.

Explore more on GitHub ↗

The tools can change.
The responsibility stays.

01

Frame the problem.

Turn unclear requirements into specific workflows, constraints and a definition of a correct result.

02

Connect the parts.

Use configuration, data and AI-assisted implementation to build something people can operate.

03

Check the result.

Test important paths and failure states. Keep the decisions, limitations and evidence visible.

Sadeq’s original digital-twin project artwork
Digital twin / interactive case studyRead case study ↗

A portfolio.
With a voice.

Ask about my projects, experience or working process. My digital twin connects a paper-cut character, speech and a maintained professional knowledge source.

Explore the full case study

The case study documents the avatar, voice and AI system. To talk with the assistant here, use “Ask Sadeq’s AI.” AI answers can be wrong—project sources remain the reference.

Sadeq Rezai.

Based in Budapest.
Curious across disciplines.

I work as a Configuration Consultant at FRAIA, between requirements, testing, data and the interface people use.

I use AI-assisted development while owning the structure, validation and integration. R is my strongest programming language; I also have practical Python and SQL knowledge.

NOW
FRAIA / Configuration
PREVIOUSLY
Peroptyx / Data analysis
LANGUAGES
Farsi · English · Hungarian · German

05 / Open to a useful conversation

Something worth
working on?

Email me

Inside the monitoring system

Peoples Clinic QA Suite / Selected implementation

The problem

Infrastructure, clinical workflows and AI-enabled features need repeatable checks. A successful page load alone does not establish that the full workflow works.

The system

  • Scheduled smoke checks for login, dashboard and logout paths.
  • UI/API regression checks covering the consultation workflow.
  • Parser and AI-behavior checks, with artifacts and screenshots for investigation.
  • Cooldown-aware alerts intended to reduce repeat notifications.

What to inspect

The public repository documents the workflows and their boundaries. These checks demonstrate selected behavior; they are not a guarantee of clinical accuracy, security or uninterrupted availability.