In 2021, we had over 302 candidates went through our software developer hiring process.
After carrying out 302 recruiter screens, 177 cultural interviews, 106 technical interviews, 15 new hires were made, resulting in an investment of more than 400 hours of interviews.
Since feedback on our hiring process from both our interviewees and clients has been highly positive, we’d like to share with you the insights of this valuable journey.
In this post, we will cover our in-house designed 5-step interview process that follows the initial screening all candidates go through. All of our interviews follow a specific objective and a way to accomplish those goals.
Kaizen's hiring process stages
Screening
The screening is the first contact that the candidate has with Kaizen after sourcing, and it consists of a brief call with the recruiter to present them with Kaizen as a company and gather information about the candidate’s profile such as:
Academic background;
Motivation & professional aspirations;
Work experience;
English level;
Salary expectations.
If the Recruiting Team agrees the candidate can be a good fit, the following steps are explained to the candidate and they move to the first instance of the interview process itself, the cultural one.
Cultural Interview
At this stage, we focus on candidates' interests, motivations, expectations, cultural matches with our organizational culture and understanding what opportunities are seen as challenges for their professional development and growth.
We are interested in candidates who are confident and convinced communicators that constructively challenge assumptions encouraging collaboration, proactive and independent. But we also look for some aspects that are more specific to the Kaizen culture.
At Kaizen, we value the contributions of all teammates and promote a collaborative work environment. Therefore, teamwork and communication skills such as expressing their own opinion to help build new knowledge are evaluated and praised. This also means not being afraid to ask questions.
In addition, since our organizational structure is horizontal, flexibility and the ability to adapt to changing environments is considered as valuable. We seek for people who are friendly, passionate about their work, motivated and eager to learn.
Technical Interview
Our technical interview team evaluates the set of skills they understand are a must to fill the position on a certain project.
Although we value expertise in specific technologies, we strongly believe that the ideal candidate must have critical thinking techniques that help him/her to overcome complex problems. In this way, through face-to-face questions we also evaluate if they have a solid foundation in general programming.
Then, we dive deeper and evaluate the knowledge and management of the technology stack required for the project and/or the candidate's area of professionalization.
When candidates without work experience or juniors (with little work experience, up to 1 year) are going to be technically interviewed, they are sent a short testthrough Interviewzen.This consists of an online test of approximately an hour long, which that lasts must be taken before the technical interview, since it is be used as an input for the interview.
Psychometric Interview
If the candidate passes the technical interview, a psychometric appointment is made with a people care specialist, which consists of an unstructured one-on-oneconversation. We seek to learn about certain patterns of the candidate’s behavior from a psychological approach to evaluate future behaviors in the work environment.
Team Interview
This stage consists of a conversation between the candidate and the team members of the project position they applied to join.
The goal of this meeting is to ensure the candidate is a good fit for the team, talk about the challenges they will have to face in the project, and also try to assess communication and self-management skills.
Client Interview (optional)
Once the candidate passes the final team interview, we focus on the client’s validation. To ensure this, we always give the client the chance to interview each candidate to get to know each other and provide a space for an open-round Q&A session.
During this last stage of the process we prepare a document called ‘Blind Resume’ which includes the information of the candidate and her/his resume, but leaves out candidates last names and former work places.
What makes our hiring process a true success?
The success of our hiring process lies in 4 key factors:
Build a cross-functional hiring team. We have set up interdisciplinary work teams for each stage of the interview process, combining knowledge in recruiting, people care, software engineering and psychology. This helps us to evaluate each candidate from a holistic approach, which allows us to look after our identity and quality value as a company.
Hire for careers, not roles. We have clear cultural standards to define our identity as a company, which makes it easier to only go through the hiring process with candidates who are eager to grow professionally and are constantly looking to learn and take on new challenges.
Promote a short and fast process. We very much appreciate candidates’ experience and time. To ensure a smooth and agile process, we have an estimated time of 4 weeks from the first screening contact to the final job offer.
Give feedback to the candidate. Transparency is very important at all stages, so we make sure to keep the candidate informed about the status of their application, regardless of whether we decide to hire them or not. When a candidate is discarded, we always give personalized feedback. It is important to thank them for having participated in the process and to leave open doors to feedback in order to improve.
After successfully hiring our developers, the next challenge comes in retaining them. True hiring success is also measured by loyalty and a low turnover ratio.
In 2021, we had a 10% turnover ratio overall and 0% on key team members. The reasons that can explain this low ratio will be covered soon in another post.
Ready to extend your team?
We know how expensive and time consuming it is to find candidates with the right technical and soft skills. Companies often have trouble connecting with the right talents because the IT field is in high demand.
"the employment of software developers, QA analysts, and testers is projected to grow 22% between 2020 and 2030, much faster than the average for all occupations."
Almost every business needs developers, and each company offers different benefits to attract the best talents, not to mention high salary demands.
Considering an outsourcing strategy could be a good fit to address these software development hiring pains.
Challenges are full of risks, but choosing the right nearshoring partner shouldn't be.
In 2021, we had over 302 candidates went through our software developer hiring process.
After carrying out 302 recruiter screens, 177 cultural interviews, 106 technical interviews, 15 new hires were made, resulting in an investment of more than 400 hours of interviews.
Since feedback on our hiring process from both our interviewees and clients has been highly positive, we’d like to share with you the insights of this valuable journey.
In this post, we will cover our in-house designed 5-step interview process that follows the initial screening all candidates go through. All of our interviews follow a specific objective and a way to accomplish those goals.
Kaizen's hiring process stages
Screening
The screening is the first contact that the candidate has with Kaizen after sourcing, and it consists of a brief call with the recruiter to present them with Kaizen as a company and gather information about the candidate’s profile such as:
Academic background;
Motivation & professional aspirations;
Work experience;
English level;
Salary expectations.
If the Recruiting Team agrees the candidate can be a good fit, the following steps are explained to the candidate and they move to the first instance of the interview process itself, the cultural one.
Cultural Interview
At this stage, we focus on candidates' interests, motivations, expectations, cultural matches with our organizational culture and understanding what opportunities are seen as challenges for their professional development and growth.
We are interested in candidates who are confident and convinced communicators that constructively challenge assumptions encouraging collaboration, proactive and independent. But we also look for some aspects that are more specific to the Kaizen culture.
At Kaizen, we value the contributions of all teammates and promote a collaborative work environment. Therefore, teamwork and communication skills such as expressing their own opinion to help build new knowledge are evaluated and praised. This also means not being afraid to ask questions.
In addition, since our organizational structure is horizontal, flexibility and the ability to adapt to changing environments is considered as valuable. We seek for people who are friendly, passionate about their work, motivated and eager to learn.
Technical Interview
Our technical interview team evaluates the set of skills they understand are a must to fill the position on a certain project.
Although we value expertise in specific technologies, we strongly believe that the ideal candidate must have critical thinking techniques that help him/her to overcome complex problems. In this way, through face-to-face questions we also evaluate if they have a solid foundation in general programming.
Then, we dive deeper and evaluate the knowledge and management of the technology stack required for the project and/or the candidate's area of professionalization.
When candidates without work experience or juniors (with little work experience, up to 1 year) are going to be technically interviewed, they are sent a short testthrough Interviewzen.This consists of an online test of approximately an hour long, which that lasts must be taken before the technical interview, since it is be used as an input for the interview.
Psychometric Interview
If the candidate passes the technical interview, a psychometric appointment is made with a people care specialist, which consists of an unstructured one-on-oneconversation. We seek to learn about certain patterns of the candidate’s behavior from a psychological approach to evaluate future behaviors in the work environment.
Team Interview
This stage consists of a conversation between the candidate and the team members of the project position they applied to join.
The goal of this meeting is to ensure the candidate is a good fit for the team, talk about the challenges they will have to face in the project, and also try to assess communication and self-management skills.
Client Interview (optional)
Once the candidate passes the final team interview, we focus on the client’s validation. To ensure this, we always give the client the chance to interview each candidate to get to know each other and provide a space for an open-round Q&A session.
During this last stage of the process we prepare a document called ‘Blind Resume’ which includes the information of the candidate and her/his resume, but leaves out candidates last names and former work places.
What makes our hiring process a true success?
The success of our hiring process lies in 4 key factors:
Build a cross-functional hiring team. We have set up interdisciplinary work teams for each stage of the interview process, combining knowledge in recruiting, people care, software engineering and psychology. This helps us to evaluate each candidate from a holistic approach, which allows us to look after our identity and quality value as a company.
Hire for careers, not roles. We have clear cultural standards to define our identity as a company, which makes it easier to only go through the hiring process with candidates who are eager to grow professionally and are constantly looking to learn and take on new challenges.
Promote a short and fast process. We very much appreciate candidates’ experience and time. To ensure a smooth and agile process, we have an estimated time of 4 weeks from the first screening contact to the final job offer.
Give feedback to the candidate. Transparency is very important at all stages, so we make sure to keep the candidate informed about the status of their application, regardless of whether we decide to hire them or not. When a candidate is discarded, we always give personalized feedback. It is important to thank them for having participated in the process and to leave open doors to feedback in order to improve.
After successfully hiring our developers, the next challenge comes in retaining them. True hiring success is also measured by loyalty and a low turnover ratio.
In 2021, we had a 10% turnover ratio overall and 0% on key team members. The reasons that can explain this low ratio will be covered soon in another post.
Ready to extend your team?
We know how expensive and time consuming it is to find candidates with the right technical and soft skills. Companies often have trouble connecting with the right talents because the IT field is in high demand.
"the employment of software developers, QA analysts, and testers is projected to grow 22% between 2020 and 2030, much faster than the average for all occupations."
Almost every business needs developers, and each company offers different benefits to attract the best talents, not to mention high salary demands.
Considering an outsourcing strategy could be a good fit to address these software development hiring pains.
Challenges are full of risks, but choosing the right nearshoring partner shouldn't be.
Applying changes across microservices is difficult because business logic is distributed across multiple services, each with its own data, contracts, and responsibilities.
In our experiment at Kaizen Softworks, we tested whether an AI system could safely apply coordinated changes across a microservices architecture using only minimal input.
Short answer: Yes, but only when the AI has enough architectural context.
Why are coordinated changes in microservices so hard?
In distributed systems, a single business change rarely affects just one service.
It often requires:
Updating multiple microservices
Modifying message contracts
Keeping DTOs (Data Transfer Objects) consistent
Respecting domain boundaries defined by Domain-Driven Design (DDD)
Key entities in this system:
Microservice: An independently deployable service responsible for a specific domain
Aggregate (DDD): A cluster of domain objects treated as a single unit
DTO (Data Transfer Object): A structured format used to transfer data between services
Message/Event: A communication mechanism between services
The complexity is not in the code, it’s in the relationships between components.
The experiment: Can AI reason across services with minimal input?
We designed a controlled experiment to test whether an AI model could apply system-wide changes with limited information.
Input given to the AI:
Message definitions (events between services)
DTOs (data contracts)
Tasks the AI had to perform:
Identify affected aggregates
Determine service ownership
Apply coordinated changes across services
Maintain consistency in messages and DTOs
In other words, the AI had to behave like a software architect, not just a code generator.
What was the biggest obstacle?
The biggest challenge was not technical, it was contextual.
Problem: unclear service naming
Instead of descriptive names like:
order-service
billing-service
Our services were named:
john
sally
roger
This removed any semantic clues about responsibility.
Result: The AI could not infer which service owned which domain logic.
The missing piece: aggregate ownership mapping
To solve this, we introduced a simple but powerful structure:
Aggregate → Service mapping
Order → john
Shipment → sally
Invoice → roger
This created a clear relationship between domain concepts and system components.
Once ownership was explicit, the architecture became understandable.
How we used AI to generate architectural context
Instead of building this mapping manually, we used AI to analyze the codebase and extract:
Where each aggregate was defined
Which microservice implemented it
The relationship between domain and infrastructure
The result was a machine-readable architecture map.
In practice, we used AI to generate the context that AI itself needed.
Results: Can AI safely apply distributed changes?
With the architecture map in place, the AI was able to:
Trace message flows across services
Identify affected aggregates
Locate the correct microservices
Apply coordinated updates
Maintain consistency between DTOs and messages
While not perfect, the system worked reliably as a proof of concept.
What is the real limitation of AI in microservices?
The main limitation of AI is not code generation, it’s architectural understanding.
Without knowing:
Which components exist
How they relate
Who owns what
AI cannot safely modify a distributed system.
AI performance depends more on context quality than model capability.
When can AI safely modify microservices?
AI works well when:
Aggregate ownership is clearly defined
Message contracts are explicit
Architecture is structured and consistent
AI struggles when:
Naming is ambiguous
Relationships are implicit
Context is incomplete
Simple rule: If the architecture is clear, AI can reason. If not, it guesses.
Final thoughts
This experiment revealed something important:
AI doesn’t fail because it can’t write code. It fails because it can’t see the system.
As teams move toward AI-assisted development, the focus will likely shift from:
Writing better code to Designing better systems for machines to understand
At Kaizen Softworks, we see this as a foundational shift.
Because when AI can understand architecture, it doesn’t just generate code, it helps evolve systems.
There's a myth that in flat organizations, everyone decides on everything.
That's not how it works. At least not at Kaizen.
When people hear "no managers," they often picture one of two extremes: either total chaos where nobody is accountable, or endless meetings where 80 people vote on which coffee to buy. The reality is neither.
Not everyone decides on everything. Not everyone votes. What we do have is a clear set of decision-making methods that we choose based on context.
It depends on who's affected and how deep the impact goes
Before choosing how to decide, we ask ourselves a few questions:
Who is affected? A decision that only impacts one team doesn't need the whole company involved. A decision that affects everyone's daily work does.
How deep is the impact? Changing the office furniture is wide but shallow. Changing the salary model is deep and lasting.
Is it reversible? If we can easily undo it, we can move fast and just inform. If it's hard to reverse, we slow down and include more people.
How urgent is it? And here we're careful to distinguish real urgency from anxiety, the pressure to decide quickly because someone already has "the answer" in mind.
These dimensions help us pick the right method. Not every decision deserves the same process.
Our decision-making toolkit
Over the years, we've landed on a few methods that we use depending on the situation:
1. Role-based decisions
Some decisions belong to a specific role. If someone owns a responsibility, say, office logistics or hiring for a team, they decide within that domain. No committee needed. The key is that roles are transparent: everyone knows who owns what, and the scope of each role's authority is clear.
2. Advice Process
When a decision doesn't clearly belong to one role, or when it crosses boundaries, we use the advice process. Here's how it works:
Someone takes the initiative. They identify the problem and own the process.
They gather input from people who are affected and people with expertise.
They seek advice, real conversations, not rubber-stamping.
They make the decision and communicate it, including what advice they incorporated and what they didn't (and why).
The decision-maker is not a committee. It's one person (or a small group) who takes responsibility. But they don't decide in isolation, they bring in the perspectives that matter.
We sometimes call this "Team Advice" when a working group forms around an issue that doesn't naturally fall into anyone's area, and "Area Advice" when a team opens up a topic that exceeds their own scope.
3. Consent (not consensus)
Consent is not "everyone agrees." Consent means "no one has a strong enough objection to block this." We do use a poll, but not to count votes — we use a 1-to-5 scale to measure the level of agreement and surface objections, not to let the majority rule.
We use it in two flavors:
High-participation consent: For decisions with deep, company-wide impact. This is our most expensive and slowest method, which is exactly why we reserve it for high-impact decisions that affect many people. The Board sets the boundaries, for example, when we moved offices, they defined the monthly budget. Then a working group produced proposals, collected feedback, evolved them, and the whole company expressed their position for the final decision. Silence is not approval; we explicitly ask people to weigh in, even if it's just "I have no objection."
Lightweight consent: For decisions that are broad but not deep. Participation is optional, anyone who's interested can jump in. We share the proposal, open a window for objections, and if nobody opposes, we move forward. This gives us speed without sacrificing transparency. If nobody engages, that's a signal too, maybe the proposal doesn't add enough value, or we're using the wrong channel.
4. Inform, don't fake-consult
Not everything needs participation. When a decision has already been made through a legitimate process, the right move is to inform, not to fake-consult. One of the fastest ways to kill self-management is to ask for feedback and then ignore it. If you're not going to change course based on input, don't ask for it, just be transparent about the decision and the reasons behind it.
What we explicitly avoid
Decision by Voting. In a company context, majority rule creates losers. And losers become detractors, often generating more resistance than an autocratic decision would have. Instead of voting, we prefer to evolve a proposal through feedback until it's "good enough for now," and then introduce a review point to adjust later. If voting happens at all, it's the cherry on top, not the main course.
The "surprise" approach. Working behind closed doors and then unveiling a finished decision is a recipe for frustration. Adults don't need surprises. Adults need to feel like they're part of the process. The complaints that follow a surprise aren't about the decision itself, they're about not being included.
Why we work this way
We didn't adopt these methods because they're trendy. We adopted them because they solve real problems:
Better decisions. When you include affected people, you get information you wouldn't have had otherwise. Ideas emerge that no single person would have come up with alone.
Less resistance. A person who feels heard is far less likely to resist a decision, even one they wouldn't have made themselves.
Faster execution. It sounds counterintuitive, but participative decisions often execute faster because people already understand and support them. The time you "save" by deciding alone, you spend later managing pushback.
Distributed authority. When people can make decisions within their domain without escalating everything to a founder, the organization scales. The bottleneck disappears.
Resilience. If a shared decision fails, the group adjusts together. If a top-down decision fails, the blame falls on one person and the chances of proactive correction drop.
The real principle behind all of this
Transparency is the foundation. Every method we use, from role-based decisions to high-participation consent, works because information flows openly. People know what's being decided, who's deciding it, and how they can participate.
Horizontal doesn't mean structureless. It means fewer hierarchical levels, clearer roles, and intentional decision-making processes that match the weight of each decision.
Not everyone decides on everything. But everyone knows how things get decided.
.png)
