Skip to content
Czanix LogoCZANIX
CZANIX
Cesar Zanis · Czanix

How I work.

Before choosing the technology, I need to understand what is happening.

A growing cloud bill, a report nobody trusts, a task that depends on copying and pasting. That is where the conversation starts. I want to understand the problem, who deals with it and what has already been tried.

From the first conversation to operations

  1. Understand the situation

    I talk to the people who use and maintain the system. I look at costs, data, integrations and team constraints. Sometimes the initial request changes once we find the cause.

    The outcome: a defined problem and a way to recognize improvement.

  2. Choose what fits

    I compare options and explain what each takes to build and maintain. Deadlines, budget and team experience matter. A solution needs to make sense after the project ends.

    The outcome: scope, priorities and the reasons behind each choice.

  3. Build and check

    I start with a part we can test in real conditions. I review it with the people who will use it, adjust what did not work and continue from there. If an assumption fails, we change direction.

    The outcome: clarity about what has been validated and what remains open.

  4. Prepare for operations

    Going live includes monitoring, documentation and knowledge transfer. The people taking over need to know how to observe, fix and extend what was delivered.

    The outcome: a way to maintain the system and clear responsibilities.

In practice, the starting point varies.

Examples of how I would approach common problems. These are illustrative situations, not client results.

The cloud bill has grown.

First I break down costs by application and look for what changed. Resizing resources may help. Sometimes the cause is a query, log retention or an architecture decision.

The team wants to use AI.

I look for a specific task with accessible data and a way to check the answer. Then I assess permissions, cost per use and what happens when the model gets it wrong.

Every report has a different number.

I go back to the data source and calculation rules. Before building another dashboard, we need to agree on what the metric measures and who maintains that definition.

Does this sound familiar?

Tell me about the context and what you need to solve. We can then see whether working together makes sense.

Talk about a projectRead my blog