I'm a Principal AI Product Leader with 15+ years building enterprise technology products — and the last several years focused entirely on shipping agentic AI systems at scale.
I've led the product development of AI systems serving 20M+ users, including a transactional banking assistant processing 28K+ secure operations per month, and a multi-agent workforce platform used by 300K+ employees. I work at the intersection of AI product strategy, agentic systems design, and enterprise delivery.
Most of what I publish here is hands-on: frameworks, prototypes, and evaluation tools built from real production experience — not tutorials.
- 🤖 Agentic systems — multi-agent architectures, orchestration frameworks, agent-to-agent coordination
- 🏦 Enterprise AI in regulated environments — banking, energy, workforce — where guardrails and governance aren't optional
- 📐 AI product frameworks — how to take agentic initiatives from process discovery to scaled rollout responsibly
- 🧪 LLM evaluation — comparing models on accuracy, latency, cost, and guardrail compliance for real enterprise use cases
- 📝 Thought leadership — writing about organizational readiness for agentic AI, model selection, and responsible deployment
| Repo | What it is |
|---|---|
| eamm | Enterprise Agentic Maturity Model — 7-level framework for organizational readiness for agentic AI |
| agent-factory-framework | End-to-end delivery framework for enterprise AI agents — 6 stages, 5 templates, decision gates |
| llm-selection-framework | Decision framework: prompting vs RAG vs fine-tuning vs agentic orchestration |
| delegation-ladder | Framework for designing AI agents that earn trust progressively — 5 levels, design mandates, delegation metrics |
| designing-for-model-uncertainty | Five design patterns for consumer AI products that earn durable trust — uncertainty signals, graceful degradation, recovery design |
| financial-agent-guardrails | Playbook for designing, implementing, and auditing guardrails in AI agents deployed in regulated banking environments |
| agentic-fraud-detection | Architecture patterns for AI agents that detect, investigate, and respond to fraud in real-time banking and payments |
| conversational-kyc-agent | End-to-end architecture for AI agents that conduct KYC onboarding conversationally — document to risk profile to approval |
| banking-assistant-anthropic | Conversational banking assistant with RAG + eval suite — Anthropic Claude API |
| llm-evals-enterprise | Claude vs GPT-4o evaluation framework — accuracy, latency, cost, guardrail compliance |
AI & Agentic Systems
Agentic Orchestration Multi-Agent Architecture LLMs RAG Prompt Engineering Guardrail Design AI Evals watsonx Orchestrate Anthropic API OpenAI API
Product
AI Product Strategy 0→1 Product Delivery Roadmap Definition OKRs & KPIs Agile / SAFe Discovery & User Research Enterprise Platform Delivery
Domains
Banking & Fintech Enterprise Workforce Regulated Environments Digital Transformation
- 💼 linkedin.com/in/aldojmacias
- 🌐 aldojm.com
- ✍️ Writing about enterprise agentic AI from production experience — The Delegation Ladder · EAMM · Designing for Model Uncertainty · Financial Agent Guardrails