We show what broke
Check-in that fell over under a crowd, matchmaking that sent everyone to the same people. The fixes are the useful part.
Case studies
Engineering stories from products we run ourselves
Eventify and ZekoHR are built, hosted and supported by the same engineers you would work with. Here is how they are put together, which AI decisions held up, and what we would do differently.
- Two live products
- Six engineering deep-dives
- No invented metrics
Why these studies
Honest
Specific
Transferable
Real architecture, real trade-offs, no borrowed logos
Client work is usually under NDA, so we write about the two platforms we own. Each study follows the same path: the context, the hard part, what we built, the architecture, the decisions that mattered and what it means for your project.
Diagrams of the mechanism
Every study has an architecture figure that shows where data flows, where the tenant boundary sits and where the model is called.
Patterns you can reuse
Multi-tenancy, offline-first clients, grounded AI answers and evaluation loops apply well beyond events and HR.
The studies
Six product engineering stories
Eventify
Building a white-label event platform
Native apps, organiser admin and a tenant-scoped API that lets one backend serve many branded events without leaks.
Read the study ZekoHRA multi-tenant HR platform with payroll you can audit
Payroll, attendance, recruitment, documents, expenses and grievances, with isolation enforced in one place.
Read the study EventifyAI matchmaking that explains itself
Embeddings for recall, rules for fairness and an LLM that only writes the reason, never the ranking.
Read the study EventifySession speech-to-text and summaries
Two-pass transcription with vocabulary pulled from the agenda, then summaries that feed search and the support bot.
Read the study EventifyA support bot grounded in each event
Retrieval filtered by tenant and event, live tools for anything time-sensitive, and a clean handoff to humans.
Read the study EventifyBadge printing and the self-service kiosk
A local-first desktop app that keeps printing and checking people in when the venue Wi-Fi gives up.
Read the study
How we work
The same discipline on your project
Every feature in these studies went through a short written spec before code, was built in two-week sprints and was tested at a live event before we called it done. We run client work the same way: a discovery sprint, a fixed statement of work where the scope allows it, and senior engineers who stay on the project.
- Written spec and risk list before the first sprint
- Working software reviewed every two weeks
- Evaluation sets for every AI feature, not just demos
- Hand-over documentation and a managed-services option
Event technology services
Apps, registration, check-in, kiosks and AI for events, built on what Eventify taught us.
ExploreAI for HR technology
Recruitment screening, policy copilots and workflow automation, informed by running ZekoHR.
ExploreAI knowledge bases (RAG)
The grounded-answer pattern behind the Eventify support bot, applied to your documents and systems.
Explore
FAQ
Frequently asked questions
Why are the case studies about your own products rather than client work?
Most client engagements are covered by confidentiality agreements, so we cannot publish their architecture. Eventify and ZekoHR are ours, which means we can show the real decisions, the trade-offs and the parts that went wrong before they went right. Client references are available on request during a scoping call.
Do you only build event and HR software?
No. These two products are where we run our own engineering in production every day, so they make the most honest case studies. The same team delivers AI services, Salesforce work and custom software across many sectors.
Can we use the same components in our product?
Often, yes. The matchmaking engine, the speech-to-text pipeline, the support bot and the kiosk are designed as services behind a tenant boundary, so they can be licensed as part of a white-label platform or rebuilt to fit your stack under an engagement model that suits you.