Skip to content

Alexandre Jaques, AI Integration Engineer. CHAOS IN. SYSTEM OUT.

Available nowLisbon, PortugalExplore the work
LLMs, APIs and automation forreal business systems.

How business chaos becomes a working system

Portrait of Alexandre Jaques
Alexandre JaquesAI Integration Engineer

business problem first

  • Customersunclear needs
  • Emailunstructured
  • CRMsiloed
  • Datascattered
  • APIsinconsistent
  1. Understand

    • Discover
    • Clarify
    • Map context
  2. Connect

    • Integrate
    • Orchestrate
    • Validate
  3. Ship

    • Automate
    • Deploy
    • Measure

Working systems

Live

  • Internal toolsfaster operations
  • AI agents24/7 execution
  • Automationsless manual work
  • Insightsbetter decisions
  • Integrationsconnected stack
01Proven in production

Proof, not promises.

The strongest work is not a gallery of screens. It is the path from a business constraint to something people can actually use.

01Live SaaS, paying customers
01Live SaaS, paying customers

Mensagi, AI customer service on WhatsApp and Instagram

A multi-tenant SaaS that answers the first contact, checks the real calendar and books the appointment. Built and operated with a partner.

Problem
Service businesses (barbershops, gyms, clinics) lose bookings after hours, and answering the same WhatsApp and Instagram questions by hand consumes the working day.
Build
pnpm and Turborepo monorepo: Next.js on Vercel, an authoritative Fastify API, BullMQ workers, one isolated container per WhatsApp number, Supabase with row level security, Redis, and LLM calls behind a prompt layer.
Outcome
In production with around twelve businesses and paying customers in Portugal and Brazil. The AI books into the real calendar, with owner approval where it is wanted and escalation when a conversation needs a person.
  • Next.js
  • Fastify
  • Supabase
  • BullMQ
  • WhatsApp API
  • LLM APIs
Open live productRead the architecture case study
02Private client system
02Private client system

Semantic search over support history

A RAG workflow that turns years of solved tickets into usable context during triage.

Problem
Resolutions lived inside old Freshservice tickets, so agents repeatedly solved familiar issues from scratch.
Build
Freshservice ingestion, Supabase pgvector, embeddings, n8n synchronization, retrieval and agent handoff.
Outcome
Past resolutions became available during triage as part of an automation program that reduced manual effort by 60%.
  • RAG
  • Freshservice
  • Supabase
  • n8n
  • LLM APIs
0315+ client deliveries
0315+ client deliveries

CRM and ERP integration delivery

Integration work across the less glamorous systems where operations either flow or stop.

Problem
Teams retyped information between HubSpot, Pipedrive, Salesforce, SAP and custom systems.
Build
Discovery, process mapping, REST APIs, webhooks, n8n, Make and Python services, from scope to handoff.
Outcome
Projects across legal, retail, healthcare, real estate and professional services reduced manual effort by 60%.
  • REST APIs
  • Webhooks
  • Python
  • n8n
  • Make
02Live MVPs

Small enough to learn. Real enough to fail.

MVPs are separated from mature production work on purpose. They prove a behavior, a market or an operating model, not a finished company.

Live, validating
Live, validating

Orçamentista, quoting for construction and ITED

A platform that takes a Portuguese construction or electrical company from client request to sent proposal, attacking the real bottleneck: waiting for supplier prices. Suppliers reply through a single link, with no account and nothing to install.

What it proved

The platform is live and the pricing engine explains every number back to the rule that produced it, but the central bet, that suppliers actually answer a cold link, is not validated yet. That is the next test, not more features.

  • FastAPI
  • Next.js
  • PostgreSQL
  • DDD
  • Explainable pricing
Open project
03Concepts

Thinking made visible.

These are product explorations, not production claims. Their value is in framing the problem, testing the flow and making a decision easier.

01
Concept, 9 screens

Padel App

A multi-role product concept connecting players, clubs and organizers through schedules, events and balanced matches.

What it proved

One shared product needs radically different priorities for each role.

  • Product design
  • Multi-role flows
  • Community
02
Concept, 8 screens

Daillo

A conversational wellbeing concept that turns a chat into a daily, inspectable supplement routine.

What it proved

AI advice only becomes useful when it resolves into clear actions and history.

  • AI product
  • Mobile UX
  • Daily workflow
04About and method

Somewhere between business and code.

I do not start with the technology. I start with what needs to happen, who is doing it today and where the process breaks.

Portrait of Alexandre JaquesAlexandre Jaques
Lisbon, Portugal

Six continuous years at ChatGuru took me from the first Customer Success hire to Director. Alongside that path, I ran IntegraNinja part-time and delivered integration work for clients. That mix of customers, operations, product and engineering is the reason I can translate between a business problem and the system required to solve it.

Based in
Lisbon, Portugal
From
Brazil
Work authorization
European Union
Availability
Immediate, full-time
Open to
Remote EU/US, Lisbon hybrid, relocation
Languages
Portuguese, native; English, full professional proficiency; Spanish, basic
The method

Chaos in. System out.

  1. 01

    Understand

    Observe the current operation and find the actual constraint.

  2. 02

    Diagnose

    Separate symptoms from the problem worth solving.

  3. 03

    Scope

    Reduce ambiguity to a measurable first system.

  4. 04

    Build

    Connect the smallest set of tools that can do the job reliably.

  5. 05

    Ship and learn

    Put it in reality, measure failure and improve the operating model.

Career

One company. Four roles. One parallel studio.

Primary trackChatGuruOct 2019 to Nov 2025
  1. Oct 2019 to 2020

    Customer Success Analyst

    First CS hire, building onboarding and support workflows from the ground up.

  2. 2020

    Customer Success Manager

    Built the team and established onboarding, retention and expansion processes.

  3. Nov 2020 to Jan 2025

    Head of Operations

    Led cross-functional teams and automation that reduced manual work by 60%.

  4. Jan 2025 to Nov 2025

    Director

    Led strategy across product, operations and engineering, including the AI chatbot roadmap.

Part-time, in parallelIntegraNinja

Founder

Jul 2021 to Jan 2025

Independent integration consultancy for 15+ clients, run part-time alongside ChatGuru.

05Everything led to this

A Oficina.

A live venture where the same method becomes public: understand a real problem, build the system, expose what breaks and teach from production evidence.

A Oficina website
Businesses do not need more AI demos. They need systems that survive the real world.

Live today

  • Public academy with free resources behind a lead form
  • Instagram comment to DM funnel, proven in production
  • Private operations panel: leads, triggers, content approval
  • Plain Postgres with numbered SQL, Docker, Coolify and systemd timers
Where it goes next

Grow the learning path and community only after the systems, content and delivery model have earned the right to scale.

  • Next.js
  • TypeScript
  • Postgres
  • Docker
  • Automation
  • Analytics
Enter A Oficina
06Lab and offline

Experiments, side quests and useful mistakes.

The Lab is where small tools, AI workflows and physical ideas can be tested without pretending every experiment is a startup.

STATUS BOARDLisbon, Portugal
01 / Building
Mensagi, A Oficina and GBrain, my own MCP knowledge server
02 / Learning
AI and Machine Learning Engineering
03 / Exploring
Model Context Protocol, retrieval quality and reliable human handoffs
04 / Based in
Lisbon
Portrait of Alexandre Jaques

Enough about work.

Outside the screen: photography, coffee, 3D printing, travel, Brazil, Portugal and making random things just to understand how they work.

07What colleagues say

Trusted when the work gets messy.

Alexandre was our very first hire at ChatGuru. He is versatile, committed, responsible, easy to work with and someone you can trust completely.
FCFernando CavalcantiCEO and Co-Founder, ChatGuru
Alex played a key role in our growth. He learns fast, adapts quickly and consistently delivered what we needed, even in the most demanding moments.
ISIaron SimisCo-Founder and CTO, ChatGuru
Have a messy problem?

I like those.

Tell me what needs to work in the real world.

Available immediately for full-time rolesBased in Lisbon, authorized to work in the EUOpen to remote EU/US, Lisbon hybrid and relocation