Product-first AI studio

AI products that ship value — not slides.

We build and run our own AI products — and partner with a select few companies to build theirs. Led by a founder who has shipped and sold real products for 20+ years.

Founder-led, senior teamEU AI Act–aware · GDPRProduction, not demos
valueonboard — deploy
$ vob deploy intellihr --prod
building candidate-relevance model ...... ok
parsing 1,284 job descriptions ........ ok
ai relevance scoring (p95 142ms) .... ok
eu ai act compliance check ......... pass
deploy → aws eu-central-1 .......... live
shipped value in 38s

Not an AI-hype shop. Product people who ship.

ÁrukeresőFounded & exited · 2004→2011
MiniCRMCRM SaaS · 2012
IntelliHRAI ATS · our flagship
20+
Years building & shipping products
1
Marketplace founded & exited
3
AI products built & run today
100%
Principals on your project
/ 01What we do

One studio, two tracks.

A hybrid model: we invest in our own AI products, and we take that same product discipline into a handful of partnerships.

ATrack A — our products

Our own AI products

We build and operate AI products end-to-end — and run them in production. We carry the risk, so we know exactly what it takes to ship and keep AI alive.

  • We fund and own the roadmap
  • Live revenue, real users
  • AWS operations, on us
BTrack B — your product

We build yours

A few high-value partnerships at a time, end-to-end: product strategy → AI build → AWS operations. You work with the principals, start to finish.

  • Product strategy & scoping
  • AI build — production-grade
  • AWS ops & cost discipline
We take on 2–3 partners at a time. Selectivity is the point.
/ 02Selected work

Production AI, in the wild.

Real systems with real users — each one stating exactly what the AI does.

intellihr.ai/candidates
Senior Backend Engineer42 candidates · AI-ranked
AI relevance
K. Nagy8y · Go, AWS96
A. Tóth6y · TypeScript91
D. Szabó5y · Python78
R. Kiss4y · Java64
M. Horváth3y · PHP52
Our flagship product

IntelliHR — the AI-powered ATS

A full applicant tracking system we built and run ourselves. AI works across the entire hiring lifecycle.

Job adsAI drafts complete, on-brand job ads from a short brief.
RelevanceCandidate–role relevance scoring ranks the shortlist automatically.
PrognosisLifecycle prognosis flags drop-off risk before it happens.
ClaudeReact / TSAWSPostgres
localhost:7000 · offline
Testületi ülés — 2026.04.18
on-device
S1Megnyitom az ülést, a napirend első pontja…
S2A költségvetési módosítást támogatom, de…
S1Szavazásra bocsátom a javaslatot.
Jegyzőkönyv
Napirend elfogadva (egyhangú).
Költségvetési módosítás: 6 igen, 1 tartózkodás.
Felelős: pénzügyi osztály — 04.30.
Privacy-first · on-prem

Offline meeting-minutes generator

Hungarian speech, transcribed and summarised entirely on-device. Built for the public sector, where recordings never leave the building.

ASROffline Hungarian speech recognition — no cloud, no upload.
DiarizationSpeaker diarization separates who said what.
MinutesA local LLM turns the transcript into ready-to-sign minutes.
Whisperpyannotelocal LLMon-prem
console.aws · eu-central-1
screening
healthy
job-parser
healthy
evaluation
healthy
12:04:01 POST /screen → 200 87ms
12:04:01 POST /parse → 200 61ms
12:04:02 evaluate(candidate=8821) ok
12:04:02 queue depth 0 · errors 0
12:04:03 autoscale 2 → 2 tasks (steady)
99.98%
uptime 90d
142ms
p95 latency
0
idle GPUs
Production AI, not demos

AI microservices on AWS

A fleet of production AI services — candidate screening, job parsing, structured evaluation — running on AWS with real SLAs, monitoring and cost control.

ScreeningAutomated candidate screening against structured criteria.
ParsingJob & CV parsing into clean, typed data.
EvaluationStructured evaluation with traceable, auditable output.
AWS ECSLambdaClaudePython
/ 03How we work

Product discipline, applied to AI.

01

Product-first

We start from the outcome and the P&L, not the model. AI is a means to revenue, never the pitch.

02

Ship fast (KISS)

The simplest thing that works, in production, beats the clever thing in a slide deck.

03

End-to-end ownership

Strategy → code → AWS operations. One team owns it from idea to uptime.

04

EU AI Act–aware

We design for compliance from day one — risk classification, transparency, human oversight.

05

GDPR & data residency

EU data residency by default. Offline / on-prem when the data can't leave the room.

06

Cost-disciplined

We size models and infra to the job. No idle GPUs, no runaway token bills.

/ 04Tech we use

We don't just call an API. We ship production systems.

Credibility through transparency — here's the actual stack behind the work above.

ModelsClaudeLLMsRAG
SpeechWhisperpyannotelocal ASR
CloudAWSECSLambdaS3
FrontendReactTypeScriptVite
Practiceon-premGDPREU AI Act
/ 05Who we are

Small, senior, and on the work.

F
The founder
Founder · Product & Engineering

Serial product founder. Founded and exited a price-comparison marketplace (2004–2011), founded a CRM SaaS (2012), and is building an AI ATS. 20+ years turning products into revenue.

G
The partner
Partner · Go-to-market

Leads partnerships and growth — translating product capability into the right few clients, and making sure each engagement actually moves the business.

You work with the principals — not a layer of juniors.

Let's build something that ships.

Tell us what you're trying to ship. If it's a fit, you'll talk to a principal within a week.

We take on only a few partners at a time.