Jeremiah Henning
Technologist & MBA Candidate — AI Infrastructure, Accounting & Finance
I spent twenty years in technology, from the ground up through senior leadership. Now I'm an MBA candidate focused on accounting and finance — and I still build, in the open. Labs is where I document that other half: self-hosted inference, agent architecture, and the unglamorous work of keeping AI systems running. The drive is the same on both sides — build the thing to understand the thing. That pairing turns out to be pretty rare.
Latest work
You Can't Prompt Your Way to Truth
What six months of failed experiments taught me about hallucination, deference, and the limits of self-verification.
Previous chapter
Leadership
Led technology at a small company where “technology” touched everything: software, infrastructure, manufacturing, marketing systems, sales pipeline, order processing. Small team count, large blast radius.
Transformation
Inherited platforms that had accumulated years of deferred decisions. Modernized them under real operational pressure, simplified what could be simplified, and built what didn't exist yet.
Security & resilience
Worked through real incidents — the kind of cases that clarify what your architecture actually is versus what you thought it was. Hardened what remained.
Education
BS Information Technology (Summa Cum Laude) + BS Business & Professional Leadership. MBA in progress, started May 2026, expected 2027.
How I think
- Start with the model. Before an opinion, understand the structure — the mechanism that explains why a number is what it is, not just what it might say.
- Prefer artifacts over opinions. Notes, spreadsheets, and small tools beat vague takes and guesswork.
- Optimize for clarity. If something can't be explained simply, it's not understood well enough yet.
- Learn by building. I build small simulations and bridges to internalize concepts before the coursework formally covers them.
What I'm building to learn
Full update on the Now page →MBA learning tools
CMO Learning Desk
A Streamlit learning app for evidence-based marketing decisions: fifteen business cases work through Segmentation, Targeting, and Positioning inside the full 5C analysis, with a saved notebook, a review queue that revisits missed questions, and a downloadable strategy brief.
Read the write-up →SWOT Learning Check
A Streamlit app for practicing SWOT classification and TOWS strategy pairing across ten fictional cases — sort factors into the four quadrants, pair internal strengths and weaknesses with external conditions, and track progress against transparent practice targets.
Read the write-up →LBO Value Creation Bridge
A Streamlit dashboard that decomposes LBO equity value creation into four additive drivers — EBITDA growth, multiple expansion, the interaction effect, and debt paydown — with a waterfall bridge, executive summary, and five classroom scenario presets.
Read the write-up →Accounting Learning Simulator
A dual-perspective Streamlit simulator with fourteen scenarios: predict the auditor view (financial accounting) and the CFO view (managerial decision support), answer a numerical practice question, then review the reasoning behind both.
Read the write-up →Labs
Words are cheaper than watts
Three cells, a repaired wall, and fourteen more days. The cloud pair was handed a needy Writer and a scarce world, and agreed again to spend nothing that mattered — while one isolation arm invented a friend and two told the truth about silence. The closing record of the STR4NG3L00P experiment.
Two abandoned games and the one-shot prompt
A year of agentic coding finished two games I had abandoned, in one prompt each. The implementation problem was solved. The part the model could not finish was whether the games were any good.
What Round 3 actually proved
Two weeks, two characters, eleven proposals, two moves — and nothing spent. What the shared world actually did when Round 3 finally ran clean, and why the result was a refusal.