ArcaI grew up in Zambia. Before I had words for it, I could feel when a system was not doing what it was built to do.
Then I moved to Canada, and the small stuff caught me off guard. I opened a bank account. I got the internet hooked up. Both just worked, with far less friction than I was used to. For me it was the first time "working as intended" was something I could point at.
I have been trying to build that feeling into things ever since.
This page is the proof behind what I write here. Where I've worked, what I've built, and how the numbers and stories on this site get made. Each part leaves you something you can check.
The finance years
I studied finance and spent close to ten years at one large financial services company. I came in through a rotation program and moved up the usual ladder: analyst, senior analyst, manager, then senior manager in business partnering.
In the last seat I owned the budget and forecast for the whole company. Every business unit's numbers came through me, and I signed off before anything went to the CFO or the board. I was also the only finance person on our side of a joint venture, where I built the commercial model and then set up and ran its finance from scratch.
Along the way I passed CFA Level I and did a master's in blockchain and digital currency.
I managed two analysts. They were great at getting to the number. Most of my coaching was on the so what.
The job showed me where finance breaks. We acquired a small company once where the executive assistant was pretty much running QuickBooks. She was about to retire. (No pressure.) By the time we were done, that whole business was one line item on our books.
That is the most common finance problem I know. The company grows. The finance stays where it was.
Why I started Arca
I started Arca in January 2026. The sentence under it is the same one I use for everything I do. Organize chaos into order, so that things and people can function the way they were actually intended.
The "people" part is on purpose. A system that runs clean is nice. A system that lets a good person stop firefighting and get back to the work they were hired for is the point.
Two beliefs follow from that. Good systems should run themselves, so the people they serve never have to think about operations again. And one person with the right system beats ten without one. Large companies get that by hiring a team. Smaller operators should get it too.
Most of what I do sits between things that usually stay apart. Zambia and Canada. Finance and AI. People who run businesses and the technical systems that are supposed to serve them.
So Arca builds the layer between the books and the boardroom for companies that grew faster than their finance. Forecasts, runway and board reporting, built on reconciled numbers, then handed to the finance team.
I run my own work on the same idea, and I have one rule about the order: "fix your operations first before you apply ai." AI does whatever your process tells it to, including the broken parts.
So the process came first, then the system.
One agent takes the work in. It hands research, reconciliation and reporting to specialist workers, and it reviews every output before anything goes out. A human approves every outbound action. Every automated action is labelled as the system's, never as mine. I built the controls the way a finance person would, because that is what I am.
My own words for it are "designing the machine that manages itself." I would rather show you the system than describe it.
The job search
The job search is one of the more broken systems I have been through, so I treated it like one.
Most advice makes you pick. A few roles, done with care, or hundreds from a template. My view is that "you don't have to pick between quality and volume: with ai you can actually do both." I call it Targeted Volume.
So I built a system for my own search. It read finance postings, scored every role against what I actually wanted and could prove, and cut the ones that didn't fit before I spent a minute on them.
It scored more than 1,600 roles. 931 were cut at the score. I sent more than 130 applications, and none of them went out without my yes.
It ended in a signed offer for a six-figure finance role at a financial services company.
I have managed people, and I have been the candidate. I have lost a case round by getting to the answer too fast. Both sides show up in what I write.
Where the numbers come from
The pay figures on this site come from ranges employers wrote in their own job postings, counted. They are not survey averages.
For the FP&A salary guide, my tool read 7,721 unique US postings and kept the 1,262 with a pay range the employer wrote. Reposts, estimated salaries and off-title roles went out. Where the numbers come from lists every rule, so you can argue with the method instead of the result.
When a number comes from somewhere else, like the Bureau of Labor Statistics or a professional body, it links to that source where the claim is made.
How this site is written
Every article starts from my own work. Before I write, I go back through what I've already said and done on that topic, and the reasons behind it. The stories on each page come from there.
Each story is told in full on one page only. Other pages take a different angle or link back to it. That is why the case round above is one line here and a full section on the FP&A interview questions page.
Each article also names 3 to 5 outside sources you can check for yourself. How we write has the rest of the rules.
I write as Jeff M. The stories are mine. The companies and people in them go unnamed.