Custom software
Web applications, internal systems, integrations, and products built from zero.
- Internal tools
- Client-facing applications
- System integrations
- Greenfield products
The right moment, engineered.
Most AI never leaves the demo. We build the kind that runs.
KairosApex designs and builds custom systems, AI agents, and process automation for companies — engineered to run, not to demo.
Four practices, one standard.
Every engagement starts the same way: understanding what your operation actually needs, before a line of code exists.
AI agents
Agents designed through to production deployment — with guardrails, tools, and observability.
- Agent architecture
- Tool and API access
- Guardrails and evaluation
- Production observability
Process automation
Mapping how the work actually happens, then automating the repetitive parts — so your people stay on the ones that need judgment.
- Workflow automation
- Internal copilots
- RAG over your knowledge base
- Document and data pipelines
Technical direction
Discovery, architecture, and the technical decisions your team needs made well.
- Discovery and scoping
- Architecture review
- Technical due diligence
- Fractional support for in-house teams
Four steps, no surprises.
- 01
Discovery
We map the problem, the constraints, and what success actually looks like. You get a written scope before anything is built.
- 02
Prototyping
The riskiest assumption gets tested first, in code. Cheap to throw away, expensive to skip.
- 03
Build
Small increments, each reviewable and shippable on its own. You see progress weekly, not at the end.
- 04
Operation
Deployed, observable, documented. The code is yours, and so is everything needed to run it.
You talk to the person who builds it.
I am Thiago Santos Joaquim. KairosApex is a lean practice by design: there is no account manager between you and the engineering, and no handoff to a junior team after the sale. When we talk, you are talking to the person who will write the code, make the architectural calls, and still be there when it is running in production.
Twelve years of backend engineering came before this — payments and financial systems at PagBank and iFood, distributed architecture at Klx. The kind of work where downtime is measured in money and a bad decision is expensive to unwind. That is the standard the consultancy is built on.
- Clean engineering
- Small functions, honest names, no dead code. Written to be read by whoever comes next.
- Small increments
- Every change reviewable on its own. No six-week branches, no big-bang releases.
- No overengineering
- The simplest thing that solves the problem in front of us. Abstractions earn their place or they do not exist.
- Your code is yours
- Full ownership, full access, complete handover. No lock-in, no proprietary layer you cannot leave.
The practice, applied to our own work.
KairosApex is a new name, not a new practice. Since 2024 the work behind it has shipped generative AI, automation, and system integration for clients across retail, healthcare, legal, and financial services. That client work stays confidential — what can be shown is below: products where the same engineering and the same AI patterns sold above are already running.
pairwork
Own product — Coming soon
Turns projects into a coordinated operation between people and AI. Agents work with secure access, real context, and complete traceability — speed without giving up human control.
kindeck
Own product — Coming soon
A domestic ERP: the operating system for a family. One private, trustworthy record of a household's operational life, with an AI assistant that is explainable and always confirmed by a person.
obrigadeirodelisboa
Client
One-page storefront for an artisanal confectionery in Lisbon. Zero-downtime deploys on Docker Swarm, CI with SAST, secret scanning and image CVE gates, complete technical and local SEO, considered accessibility, cookieless analytics.
Smaller projects, same standard.
Sites, one-off automations, and MVPs. Not every problem needs a consulting engagement — some need one person who knows what they are doing, for a few weeks.
Before you book.
- How does an engagement start?
- With a call, then a discovery. The discovery ends in a development roadmap: the scope broken down, the hours estimated, and a closed price for the build. That document is yours — if you take it elsewhere, that is a legitimate outcome.
- Who owns the code?
- You do, completely. Repository, infrastructure, documentation, credentials. There is no proprietary layer and nothing you need me for after handover.
- What does a project cost?
- It depends on scope, which is why discovery comes first. There is no price list here because a number without a scope is a guess.
- And if the scope changes?
- Fixed price against fixed scope. When the scope changes mid-project — and sometimes it should — we re-estimate before the work happens, never after it. No invoice arrives as a surprise.
- How long does it take?
- A focused automation, weeks. A production application, a few months. You get a schedule at the end of discovery and weekly visibility against it after that.
- What stack do you work in?
- Java with Spring on the backend — that is where twelve years and most of the payments and fintech work sit. Python for AI and automation, TypeScript and Next.js on the web, around Kafka, Kubernetes, Docker and Postgres. The list is longer than that, but it is rarely the interesting question: what matters is whether the stack fits the problem, and you will hear it in discovery if yours does not.
- Where are you based?
- Portugal. Engagements run remotely across Europe and the Americas, in English or Portuguese — and when a project genuinely needs someone on site, international travel is on the table.
The right moment is usually earlier than it feels.
A conversation costs you thirty minutes and commits you to nothing. Bring the problem, even if it is still vague.