06 / COMPANY NOVI SAD / SERBIA

SMALL TEAMS.BIGSYSTEMS.

PROJECT023 is a compact technology company combining intelligence, product thinking and engineering for difficult systems, private engagements and experimental research.

PROJECT023 DJVD DOO
FOUNDED 2024
NOVI SAD / SERBIA

01 / PROJECT023

INTELLIGENCE
+
ENGINEERING.

Project023 works across AI systems, software engineering, data intelligence and digital product development.

Our operating idea is simple: keep strategy, system design and engineering close enough that difficult work does not disappear between handoffs.

We prefer small teams with direct ownership over large layers of coordination.

Legal namePROJECT023
DJVD DOO
Founded2024
BaseNovi Sad
Serbia
FocusAI / Software
Data / Product

02 / OPERATING MODEL

CLOSE TO
THE
WORK.

Complex systems become fragile when responsibility is fragmented across too many disconnected roles.

We keep the problem, the architecture, the product experience and implementation connected through the engagement.

01 / DEFINE

Make the real problem visible.

Clarify the decision, workflow, constraints, risk and success conditions before architecture hardens.

02 / BUILD

Turn intent into a system.

Design the architecture, interfaces and intelligence layer together with implementation.

03 / VALIDATE

Test behavior, not slides.

Execute the system, inspect failure modes, correct assumptions and reduce uncertainty.

04 / OPERATE

Ship something people can use.

The destination is a working capability integrated into real operations or product use.

03 / PRINCIPLES

HOW WE
CHOOSE
TO WORK.

These principles are deliberately practical. They shape how we describe work, how we protect private engagements and how we separate production claims from research.

01

Strategy stays close to engineering.

System decisions remain connected to business intent and implementation reality.

02

Confidentiality is part of delivery.

Client identity, internal data and proprietary implementation remain private when protected by NDA.

03

Production matters.

The goal is not a presentation of technology. The goal is a working capability people can actually use.

04

Research stays honest.

Experimental non-LLM work is presented as research until evidence supports stronger production claims.

04 / CAPABILITY SYSTEM

ONE TEAM.
FOUR
FORCES.

AI, software, data and digital product work are connected disciplines, not separate departments passing work across a wall.

05 / NEXT

READY TO
BUILD
SOMETHING?

Start with the problem, the desired outcome and the constraints. Proprietary details can stay private until the right confidentiality boundary is in place.