Integrations, data pipelines, and AI-assisted automation. The boring stuff that matters, where a silent failure means wrong numbers on a dashboard, a missed compliance step, or a broken client experience.
Most companies outgrow their tools before they realize it. Processes that once ran on spreadsheets and manual handoffs now span five platforms, three teams, and dozens of automations that were never designed to work together.
I'm the person who learns how the business actually operates, designs the system it needs, and builds it. Not just the happy path: the failure modes, the edge cases, the "what happens at 2am when nobody's watching" scenarios.
I map how a business really works, not how someone documented it two years ago, and design the operational system it needs. Data models, lifecycle states, sources of truth, failure handling, the full picture.
CRM to invoicing. Property management system to reporting. Compliance screening to investor onboarding. I connect systems that were never designed to talk to each other, and make sure they keep talking.
LLM document extraction inside production workflows, with the guardrails that make its output safe to write into a system of record: measured against real data, conflicts reported instead of guessed, every run logged.
I build automations that handle real consequences: money, compliance, client experience. That means monitoring, error handling, documentation that stays true, and someone watching whether it's actually running.
These are real projects. Real constraints. Real consequences for getting it wrong.
Everyone blamed the AI for missing rent figures. The real cause was a file conversion that deleted every tracked change before the model ever saw the document. How I found it, and the pipeline I rebuilt around what it taught me.
Consolidating a property portfolio's operational data into a single reporting base, and the sync forensics that explained why a third of active leases showed $0 rent.
Transforming a manual compliance process into an automated pipeline that screens investors against global sanctions databases (OFAC, FinCEN, SEC) before a single dollar is accepted.
When your automation platforms can't tell you they've stopped working, you need something outside the blast radius watching on your behalf. I designed and built a centralized monitoring system that runs 24/7 at zero cost.
How I think about the decisions behind the systems. Reusable thinking tools from real projects.
Docs for a live system start going stale the day they're written. Six files, a handful of rules, and the reasoning that keeps a system's documentation honest across months of changes, many people, and many AI sessions.
Most technical recommendations die in meetings, not because they're wrong, but because they were never written down. A practical guide to the document that changes that.
The tool is never the point, the process is. But for the record:
If your team's operations run on interconnected systems and you need someone who can own them end to end, I'd like to hear about it.