Ad tech development
Ad tech built to order, by people who run it in production.
Ad servers, Prebid bidders, contextual engines and the data pipelines behind them.
- Reference
- Engerio, Slovak ad network
- In production
- since 2022, 100k req/min peak
- Stack
- Python, PostgreSQL, React, AWS
- Team
- 10 engineers, Prague
Capabilities
What we build
01
Ad servers and decisioning
Campaign selection by context and dynamic CPM, answering bid requests in ~100 ms.
Engerio · chapter 2 →02
Prebid integration, adapters and modules
Prebid.js bidder adapters that pass review, Prebid-compatible bid servers, wrapper setup.
Engerio · chapter 10 →03
Contextual targeting engines
Real-time page classification that matches ads to content, with no user tracking.
Engerio · chapter 3 →04
Political-ad compliance (TTPA)
Sponsor declarations, labels and transparency notices for political ads.
Engerio · chapter 5 →05
Native and rich formats, loaded only when seen
Lightweight widgets and formats that load when they come into view.
Engerio · chapter 6 →06
Reporting and analytics
Campaign and yield reporting in one interface for publishers and advertisers.
Engerio · chapter 8 →07
Third-party ad server integration (Equativ)
Extra demand in allotted slots next to your own campaigns.
Engerio · chapter 9 →08
Content scraping and data pipelines
Scrapy pipelines that keep content from many publishers fresh and structured.
Engerio · chapter 4 →
Honest advice
Your own ad tech, or an off-the-shelf platform?
Building custom isn't always the right call. This is how we decide with you.
Use an existing platform when
- Standard formats are enough
- Your team is small
- An off-the-shelf ad server covers your demand
Build your own when
- Platform limits block revenue
- Revenue sharing costs more than running your own team
- You need your own data and contextual targeting
- TTPA compliance requires a level of control you can't get from an off-the-shelf platform
Process
How we work
Discovery
What the system must do, and what it may cost.
Architecture
Components, data flow, scaling and cost model.
Iterative build
Working increments every 2 weeks.
Launch
Staged rollout with monitoring from day one.
Operate
We run it, or hand it over with documentation.
Proof
Engerio · Ad server
~100 ms
Campaigns compete inside Engerio, then its best bid competes with other bidders in the publisher's Prebid auction — in ~100 ms.
Read the chapter →Engerio · Contextual engine
No tracking
No user tracking. Ads are matched to page content in real time.
Read the chapter →Engerio · Scale & cost
100k
requests per minute at peak, on AWS across multiple availability zones.
Read the chapter →iMeteo · In-app monetisation
1 month
Prebid with Google Ads and Adform in a Flutter app, plus a consent-or-pay subscription.
Read the case →
Our stack
- Prebid.js
- Google Ad Manager
- Equativ
- Adform
- AWS
- TCF v2.2
- TTPA
- Python
- React.js/Typescript
- Scrapy
- PostgreSQL
- Prebid-compatible bid server (Python)
- Cloudflare
- Grafana
- Celery
- Redis
FAQ
Before you build
How do you price a build?
A fixed price agreed after discovery. Discovery defines scope, architecture and cost before you commit.
Who owns the code?
You do, including documentation. We can keep running it or hand it over.
How long does a build take?
It depends on scope; we ship working increments every 2 weeks. The iMeteo ad stack took 1 month.
Can you work with our developers?
Yes — alongside your team or as your entire development team.
Do you sign an NDA?
Yes, before any access.
Talk to an engineer
Tell us what you need to build. We reply within 24 hours.