Michael

Builder and implementation

If our number differs from the controlling formula or source, we investigate and explain the difference.

Michael leads implementation and test review. He previously worked at Deloitte and now works in data and AI technology, bringing a consultant's systems perspective to the formulas, edge cases, and shared runtime behavior that keep PennyCalc's tools consistent. His review areas include debt payoff, auto loan, mortgage-rate history, payroll and minimum-wage history, and the methodology that documents sourcing and verification.

His standard is that a calculator should document why it may differ from a reader's lender, paycheck, or statement. Where an authority specifies an intermediate rounding rule, the implementation and regression fixture should follow it; where real-world providers use additional inputs, the limitation should be visible instead of disguised as false precision.

Michael's perspective is that most of the bad numbers on the internet are not malicious - they are someone who built a calculator in an afternoon, didn't model the edge case, and shipped it. The fix is patience: build the test fixture before the calculator, verify against the public source, then ship.

AI tools may assist with drafts, implementation, and test scaffolding. Michael is named here because he owns the human implementation and test review of the tools assigned to him. See the review process.

pages with this reviewer byline

topic areas covered

Lender disclosures · IRS · FRED · BLS

primary sources

Test fixture first

build pattern

What Michael covers

Debt payoff toolsAuto loansMortgage rate historyMethodology documentationPayroll-tax implementation

My editorial standards

Match documented rounding rules. If the controlling publication rounds an intermediate figure at a specific step, the implementation and test fixture should do the same. Provider-specific fees or inputs that PennyCalc does not collect are called out as limitations.
Model the edge case before the headline path. Most calculators handle the typical user; the design starts with the atypical user (negative trade-in equity, an extra principal payment in month one, a self-only HSA that switches to family mid-year) and lets the typical user fall out naturally.
Cite the controlling source. Every material external rate, limit, schedule, or rule needs a nearby source link or a clearly linked reference page before it ships.
Update the test fixtures when the underlying authority updates. New IRS Pub 15-T issued? The withholding tests get re-baselined before the calculator output changes. The order matters.
Document the rounding rule next to the formula. Inline. Not in a separate doc. The next maintainer reads the file, not the docs.

How I verify a fact before it ships

Lender-facing math (mortgage, auto, debt): replicated from the Truth-in-Lending disclosure formulas, cross-checked against a published lender's Loan Estimate sample.
Withholding math: IRS Publication 15-T, percentage method, current revision - tested against several known salary scenarios published in payroll-industry calculators.
Historical rate series (mortgage, auto, payroll): FRED series ID specified in the methodology, refreshed annually.
Minimum wage and payroll-tax-rate history: Department of Labor and SSA primary sources, cross-checked against the BLS handbook tables.

Signature work

Methodology

Documents PennyCalc's source hierarchy, update cadence, AI-assisted workflow, review roles, testing expectations, and correction process.

Auto Loan Calculator

Handles positive or negative trade-in equity, exposes taxes and fees as inputs, and publishes the amortization and total-interest math. Pairs with the auto-loan-rate history page.

Debt Payoff Calculator

The avalanche-vs-snowball comparison that runs both strategies in parallel against the same debt portfolio and shows the actual interest-paid difference. The right answer is usually whichever the reader will actually stick with.

Federal Minimum Wage History (1938-2026)

The long-form historical reference with the legislative context, currently one of the higher-cited PennyCalc pages in Microsoft Copilot grounding queries.

Pages reviewed by Michael

Working with the team

Michael owns the human implementation and test review. Josh reviews financial assumptions and data; Jessie reviews the explanation and sourcing presentation. AI may assist with code and test scaffolding, but passing generated tests is not treated as sufficient release approval.

What I'm working on now

Michael's current review priority is calculator-state privacy, runtime regression coverage, accessibility, and build checks that catch structural errors before a release reaches production.

Spotted an error?

A mismatch with the controlling formula, publication, or a documented real-world statement is a high-priority report. Use the corrections page with the page URL, inputs, observed result, expected result, and source.

PennyCalc's named reviewers are accountable for the financial, editorial, and implementation work assigned to them. Read more on the about page and in the methodology.