Engineering leadership · AI systems · enterprise architecture

I lead teams that turn AI into reliable systems.

I'm Oscar Ocampo, an engineering manager and hands-on architect with 18+ years of experience. I build the teams, platforms, and operating practices that move AI from promising experiments into secure, useful, measurable work.

Read my thinking Explore selected systems

Anchorage, Alaska · Remote U.S. leadership

18+years building enterprise systems

35+corporate systems connected through data and analytics

5,000+people supported through enterprise platforms

$500K+annual recurring savings from systems my teams shipped

Selected systems

Platforms, practices, and products that changed how work gets done.

The problems I led, the systems I designed, and the outcomes they created.

AI-first delivery system

Cronus

An AI-first delivery model that puts specialized agents at the four points where a team loses time: interpreting requests, collecting status, preparing the daily meeting, and reconstructing the week.

Keeps Azure DevOps as the system of record instead of creating a parallel process
Four narrow agents draft, brief, plan, and explain—none of them writes to Azure DevOps
Turns delivery and quality signals into continuous-improvement insight

AI-first engineeringAgent orchestrationDeveloper experienceDelivery intelligence

Enterprise AI platform

Agent platform & enablement

A secure enterprise platform that progressed from conversational AI into retrieval, document analysis, specialized assistants, agentic workflows, and reusable tool integrations.

Sustained production use across enterprise functions, not a pilot that stalled
Established reusable RAG, MCP, Agent Skills, search, and tool-use patterns
Paired adoption with governance, security, privacy, and auditability controls

RAGMCPAgent SkillsResponsible AI

Knowledge infrastructure

Document intelligence & semantic retrieval

A knowledge platform that converts complex historical documents into searchable, traceable answers through validation, OCR, indexing, vectorization, and semantic retrieval.

Reduced difficult document research from manual searching to seconds
Combined structured validation with AI-assisted retrieval
Designed for enterprise permissions, reliability, and supportability

Document intelligenceVector searchOCRKnowledge systems

Engineering enablement

AI-assisted quality loop

A practical developer workflow using AI coding assistance, pull-request agents, and automated browser testing to shorten feedback loops without removing engineering accountability.

Reviews pull requests for quality, security risk, and technical debt
Uses AI-assisted Playwright testing to increase validation coverage
Keeps human ownership explicit at design, review, and release boundaries

GitHub CopilotCode review agentsPlaywrightHuman oversight

How I work

AI is a systems discipline.

Start with the outcome

Define the decision or result the system must improve before choosing a model, framework, or interface.

Treat context as architecture

Give agents deliberate context, constrained tools, clear permissions, and explicit workflow boundaries.

Make trust operational

Build evaluation, observability, security, human review, and change control into the delivery system.

Design for adoption

A capable platform only creates value when teams can understand it, use it safely, and improve it together.

Writing

Notes on building software with AI.

Practical essays about engineering organizations, agent foundations, delivery systems, and the operating model required to make AI dependable.

Agentic software6 min read

AI Is Not Eliminating Software Engineering

Agentic software moves the engineering question from what to build to what a machine may be allowed to decide, and what it must never be allowed to change.

Read article

Engineering leadership4 min read

Cronus: Making AI-First Delivery Repeatable

An AI-first delivery model in production: four narrow agents that draft work items, brief the daily meeting, plan the week, and explain what actually happened.

Read article

AI platforms4 min read

AI Agents Need an Operating Model, Not Just a Model

Five predictable failures explain why an impressive model is not enough to create a trustworthy production agent.

Read article

Experience

Engineering leadership with enough technical depth to make the hard decisions.

I build and lead multidisciplinary engineering teams spanning software, data, and AI. My scope includes hiring, coaching, performance, technical strategy, architecture, roadmaps, and end-to-end delivery. I stay close to design and implementation where my involvement creates leverage, while giving engineers meaningful ownership.

My work spans enterprise AI, cloud platforms, developer tooling, document intelligence, data systems, digital workplaces, and business-critical integrations. I have led remote product engineering and delivered systems across North and South America.

Engineering managementAI platform strategyAgentic workflowsMCP & tool integrationRAG & vector retrievalResponsible AIDeveloper enablementCloud architectureDocument intelligenceData platformsEnterprise integrationProduct delivery

Let's connect

Building an AI platform, an engineering team, or a better delivery system?

oocampo@outlook.com +1 907 519-2310