BKStart a conversation ↗
Independent consultant · Ireland

BrianKeating

AI Systems Architect

Designing multi-agent, multi-model systems for complex, high-consequence work.

Architecture with range.From C++ systems to AI agents.Making complexity useful.Built for the real world.
Scroll to enter →

Journey / Range

A career built close to the system.

  1. 01

    C++ / Banking systems

    Learning that reliability is not a feature you add later.

  2. 02

    Robotic factory automation

    Software connected to physical machines, feedback loops, and real operational consequences.

  3. 03

    Telecoms / distributed systems

    Working at network scale: services, integration, resilience, and observability.

  4. 04

    International energy projects

    Consulting across complex commodity and energy environments where technical decisions shape commercial outcomes.

  5. 05

    AI systems architecture

    Bringing multi-agent, multi-model workflows into practical, robust delivery.

Working surface
  • C++
  • C#
  • .NET
  • Java
  • TypeScript
  • Angular
  • Python
  • AWS
  • Azure
  • Kubernetes
  • Terraform
  • Data platforms
  • AI systems

Now / Orchestration

Multi-agent systems, built to make better decisions.

I design and deliver AI systems that bring the right models, tools, context, and evaluation together — turning complex operational and market information into useful, dependable outcomes.

International energy consulting → AI-native delivery

SignalhistoryresearchdomainContextplantoolscriticAgent teamEvaluationDecision

Open source / Signal

Zeus: rebuilding a specialist tool for the modern radio world.

I founded and grew Zeus into a modern SDR application for the OpenHPSDR ecosystem — moving a specialist category beyond its 1990s desktop roots toward a progressive, multi-platform, multi-device experience.

Zeus is used across the amateur-radio community and recognised by Apache Labs, a flagship provider of OpenHPSDR hardware. The work is as much about preserving a serious technical platform as it is about making it feel alive for its next generation of users.

  • Modern SDR experience
  • Multi-platform, multi-device
  • Open source, built in public

Explore on GitHub ↗

Commodity intelligence / In progress

From market noise to considered analysis.

A robust AI analysis harness for commodities: bringing together signals, research, specialist context, and disciplined evaluation to produce deeper, more dependable insight.

  • Multiple signals
  • Specialist context
  • Robust evaluation
  • Better decisions

Contact

Have a difficult system to make useful?Let’s start with the signal.

Start a conversation ↗Read the writing →