01
Agents and chatbots
One task, a few tools, and a person where the answer matters. Classify, search documents, or leave a draft. Not a generic chat so you can say there is AI.
Brickeer Consulting
We make things happen.
End to end, with quality, and with documentation someone else can read.
An engagement with scope, milestones, and a close. Three practices: data, agents, and data science. It opens with a sponsor and a written scope. It closes when someone on your team operates it.
- Agents
- Chatbots
- Airflow
- Dagster
- Databricks
- BigQuery
- Spark
Practices
What is hired
02
Pipelines that can be run again
Airflow or Dagster hold the schedule, the retries, and the alert when something fails. The pipeline can be started again without opening the laptop of the person who wrote it.
03
Where the data already lives
Databricks, BigQuery, or Spark, depending on the volume and the account you already use. The process stays on that platform. It does not move so the demo looks better.
The consultant
Carlos GomezConsultant · Madrid
I sign the engagements at Brickeer Consulting. Agents, chatbots, and pipelines, from the first scope until someone on the team operates them.
The flow
Initiation, plan, execution, control, and close. Each phase leaves something that can be read or run.
- 01
Initiation
There is a sponsor, a decision, and a success criterion. If one of the three is missing, the engagement does not start.
Remains: A one-page charter.
- 02
Plan
Scope, milestones, risks, and what stays out. It fits on two pages. Anything that changes later is written down.
Remains: A scope plan.
- 03
Execution
The work is built by milestone. The pipeline, the agent, or the analysis stays in a repository, and each milestone can be run.
Remains: A delivery that runs, with the milestone closed.
- 04
Control
A short report: done, next, and risk. The sponsor decides whether to continue, cut, or stop.
Remains: A one-page status report.
- 05
Close
Acceptance and handover. The engagement ends when someone on the client's team operates it and can see a failure.
Remains: A close note: owner, failure, and acceptance.
Cases
Two engagements in operation
- Data engineeringContaineerReal estate credit intelligence. Property, transactions, and debt, in the United States and the United Kingdom.DeliveryThe engagement left a platform that brings that data together and can be queried again. Not a map built for a demo.In operationIt is published, and people use it to look at maturities, deterioration, and deals. The client's team operates it.containeer.com
- AI applicationsNorthaA birth chart in plain language. An agent answers about the chart of the person who asks, and the reports can be read.DeliveryThe agent starts from that chart, not from a generic chat. The report comes out and can be kept. It does not predict events.In operationIt is running: a person creates a chart, asks, and opens the report. The product stays published.northa.center
If there is already a decision, a sponsor, and an idea of what close looks like, the engagement can open.
Open an engagement