JoséZechel
PT Contact

Data and AI Consultant

José Zechel

Data and AI Specialist - Founder of Brasys

I help companies solve real problems with Data and Artificial Intelligence: the problems that cost money, block operations and keep teams awake. The focus is production-grade solutions, not demos.

Recent work includes AI agents reaching ~90% support efficiency, RAG and RLM projects with PGVector and GCP, migration of ~80 Oracle RAC environments to PostgreSQL on AWS (~100TB) at MercuryGate, and data governance for corporate AI projects.

Contact me View portfolio
18+

Years with databases

80+

Migrated environments

100TB+

Managed data

PT/EN

Languages

services

Four tracks for solving real data problems

I work where data, infrastructure and AI need to leave the slide deck and run in production.

PostgreSQL and databases

Diagnostics, tuning, migrations and architecture for critical environments across Oracle, PostgreSQL, MySQL, MongoDB and SQL Server.

  • Oracle RAC to PostgreSQL migrations
  • PostgreSQL upgrades from 9.2 to 17
  • Performance, backup, HA and troubleshooting

BI and Analytics

Modeling, pipelines and analytical layers that turn fragmented data into reliable decisions.

  • Metric governance
  • Dashboards and analytical datasets
  • Ingestion and update pipelines

Data Governance and LGPD/GDPR

Pragmatic governance, permissions and traceability for corporate data and AI projects.

  • Source and access mapping
  • Retention and consent policies
  • Governance for RAG and internal knowledge

Agentic Engineering

AI agents with state machines, guardrails, auditability and integration with real channels such as WhatsApp, web chat and internal APIs.

  • RAG, RLM and PGVector
  • LangGraph, Gemini, GPT-4o and Claude
  • Business system automation

Real case - database migration at scale

How we migrated 80 Oracle RAC environments to PostgreSQL on AWS

~100TB combined, minimal downtime and thousands of dollars saved in Oracle RAC licenses.

  1. 01

    Oracle RAC on AWS RDS

    ~80 environments

  2. 02

    Compatibility analysis

    ora2pg, custom scripts, AWS SCT, AWS DMS

  3. 03

    Schema conversion

    PL/SQL to PL/pgSQL with automation and manual review

  4. 04

    Replication and cutover

    AWS DMS with controlled windows

  5. 05

    PostgreSQL on AWS RDS

    Multi-region target

Minimal downtime per environment - Functional parity validation before cutover - POCs and processes documented for the internal team

portfolio

Projects where operations mattered more than the demo

Selected cases focused on scale, availability, governance and measurable return.

MercuryGate - International logistics (TMS) Completed

Migration of ~80 Oracle RAC environments to PostgreSQL RDS on AWS

Work on a large-scale migration from Oracle RAC to open-source PostgreSQL on AWS RDS, with minimal downtime per environment. Scope included migration strategies, POCs, training, validation processes and documentation for the internal team.

  • ~80 Oracle RAC environments migrated
  • ~100TB+ of combined data
  • Thousands of dollars saved in annual Oracle licenses
  • Minimal downtime per cutover
Oracle RACPostgreSQL RDSAWS DMS/SCTora2pgPL/SQL -> PL/pgSQL
Multiple clients - Brazil In production

AI support agents at scale

Design and deployment of conversational AI agents for companies that need to scale support without losing quality, with real guardrails, internal integrations and full auditability.

  • ~90% support efficiency gain
  • Legacy and proprietary API integration
  • Auditable state machines
  • Guardrails for critical calculations
PythonLangGraphGeminiGPT-4oClaudePGVectorWhatsApp
Companies with internal knowledge bases In production

RAG and RLM for corporate knowledge access

RAG and RLM projects that give teams natural and secure access to documents, policies, contracts, technical bases and internal knowledge usually buried in folders and systems.

  • Semantic indexing of heterogeneous corporate sources
  • Data governance with team-level permissions
  • Deployment on Google Cloud Platform
  • Continuous ingestion pipelines
PGVectorGoogle Cloud PlatformLangChainLangGraphVertex AI

courses

Technical content for people who need to execute

YOUTUBEFREEPT-BR

PostgreSQL from Zero

A complete PostgreSQL course for people who want to master one of the most respected open-source databases.

PostgreSQL 16, advanced SQL, performance, backup/restore

2K+ students - 40+ lessons - published in 2024 Watch on YouTube
IN PRODUCTION2026

AI Agents in Practice

Building AI agents from scratch to production deployment, focused on real Brazilian business cases with Gemini, OpenAI and Claude.

Python, LangChain, LangGraph, Gemini, GPT-4o, Claude

Expected launch: Q3 2026
IN PRODUCTION2026

Automation with Make and n8n

From basics to advanced no-code/low-code automation for Brazilian professionals, with real cases and API integrations.

Make.com, n8n, webhooks, Brazilian APIs

Expected launch: Q4 2026

technical depth

A stack with years in operations, not just labs

Databases

Oracle (incl. RAC, TDE) 12+
PostgreSQL (9.2 -> 17) 10+
MongoDB 5
SQL Server - MySQL 6

Migrations and Cloud

AWS (RDS, DMS, EC2) 8
Google Cloud (Vertex AI) 3
Oracle Cloud (OCI) 5
Docker - Linux ops 8

contact

Does your company have a real problem?

I work with companies from any industry facing real problems involving Data and Artificial Intelligence, from bottlenecks that block growth to inefficiencies that limit operations. I work through Brasys, my consulting company, with one-off or recurring projects in Portuguese or English.

Schedule 30 minutes WhatsApp