We turn complex product and operational problems into software people can rely on. We make clear decisions, engineer carefully and build for day-to-day use.
A few systems I’ve built across cloud platforms, workflow software and enterprise operations.
I redesigned a multi-component document-processing platform for Google Cloud, replacing one-off environment deployments with a repeatable delivery model.
I built a focused service and microfrontend to manage complex archival sessions, packages and workflow-driven processing inside a larger enterprise platform.
I worked across the database, backend and frontend of a business-critical retail ERP, improving performance, integrations and day-to-day reliability.
We stay with the work from the first technical decision to production. Product choices stay close to the code, so you do not have to repeat the context at every hand-off.
We work out what is worth building, then ship a version people can use.
We make infrastructure repeatable, observable and easier to change.
We automate repetitive work and connect systems where it saves real time.
You see working software early. We explain the trade-offs, share the decisions and make the next step clear.
We agree on the users, constraints and business value before we choose the technology.
We turn the riskiest assumption into working software early.
We plan deployment, observability and maintenance from the start.
We use real behaviour and feedback to decide what to improve next.
Working principles / P01—P04
■ Make it useful first.
■ Keep it clear.
■ Build. Test. Improve.
■ Use technology for a reason.
I founded Ejupi Labs and lead its engineering work from Switzerland.
You work directly with me on architecture and delivery. That keeps decisions fast, trade-offs clear and context intact from our first conversation to production.
I draw on product engineering, cloud infrastructure, DevOps, workflow orchestration, enterprise systems and applied AI. I use the technology that solves the business problem. The stack is never the point.
I use Labs to test ideas in public. Each project tackles a real workflow, a technical limit or a better way to get work done.
CareerOS keeps verified career history, evidence-backed resumes, job matches and applications in one private desktop workspace. Optional AI runs on the device.
DjenisAiAgent is a permission-aware Windows automation experiment. It turns plain-language requests into visible, inspectable actions.
DIG is a real Gopher client with a browser demo that makes requests, selectors and responses easy to inspect.
IntegraDraw turns numerical integration into something you can see: change the function, method and interval, then watch the estimate converge.
ELIZA Lab is a small, local rule-based dialogue experiment. Every response shows the rule behind it; nothing leaves the browser.
VECTOR gives placement coordinators one clear view of the whole cohort. Search students and hosts, filter by status and move milestones forward; every change stays in the browser.
Start with the work, not the model. A useful AI feature should help with a real decision, bottleneck or repeated task. You still need to understand the inputs, check the outputs and plan for failure.
I map the workflow first, then find the judgement worth supporting and define a result we can measure. Only then do I choose the smallest system that can do the job reliably.
Building a new product? Has a system hit its limit? Is a workflow taking too much time? Tell us what is happening and we’ll suggest a practical next step.
CONTACT / SWITZERLAND
Your message goes straight to Ejupi Labs. We will tell you plainly whether we can help and, if so, what a useful first conversation should cover.
info@ejupilabs.com