IT Portfolio

Business logic doesn't care what you're selling.

I embed inside a customer's actual environment and build the system underneath the relationship — data models, dashboards, and AI-assisted tooling, not just a platform login. Proven end to end across a single fictional company's entire arc: plant floor, boardroom, and the deployment cycle a vendor runs to win the account.

Read the full story

For two and a half years I was the technical point of contact embedded with enterprise accounts — a multi-state banking client, a company's first LATAM enterprise portfolio — running executive QBRs, reconciling revenue data in Salesforce, and translating technical risk into business narratives for SOC managers, IT directors, and executives. The work was always deployment work. The title just said "Technical Success Manager."

Every platform I worked in had a ceiling — a UI that showed me the output of a data model I couldn't see, let alone build. So I stepped back and built the layer underneath: the dbt models, the health-score logic, the alert pipeline, the QBR — from scratch, not from a vendor's dashboard. Then I kept going, because the same pattern holds everywhere: a churn signal, a P&L, and a conveyor belt sensor all answer the same question — is this healthy, and what happens if we don't act? I built one fictional company's full arc to prove it: plant-floor IoT, the BI layer underneath it, a board-level expansion case, and the pre-sales deployment cycle a vendor would run to earn that company's trust. AI shows up throughout as working infrastructure, not a feature — a correlation engine in the deployment cycle, anomaly triage on the plant floor, a thinking partner while I stress-tested the harder technical calls. Pawsome Provisions is fictional. The systems aren't.

Core Competencies

Four stages of one company's growth — from the plant floor to a boardroom decision. Explore them in any order, but they were built to be read left to right.

Plant Floor

Where it started — sensor data, predictive maintenance, and a monitoring tool deployed for a plant manager, not a data engineer.

Python · Streamlit · Pandas · NumPy

View Project

BI & SQL

Turning floor and sales data into a single source of truth — schema design, analytical SQL, a self-hosted BI stack built to run on whatever infrastructure is actually there.

PostgreSQL · Docker · Metabase · SQL

View Dashboards

Customer Health

The pivot — rebuilding, from the data up, the account health and revenue system I used to operate through a vendor's UI.

View Customer Health Stack

Expansion

Where both worlds collide: should this manufacturer take on SaaS-style contracts? The board-level case, built end to end.

The real question: idle capacity from over-investing during growth — does private-label capacity pay for itself, or is it good money chasing a sunk cost?

View the Business Case

Two Deeper Threads

The core arc above is the whole story if you want it to be. These two go further, into two different kinds of judgment: an internal build-vs-buy decision, and a vendor's own pre-sales cycle.

Acquisition

Should Pawsome buy a veterinary SaaS platform? The build-vs-buy judgment, applied to Pawsome's own decision-making — not a vendor's pitch, an internal one.

Build vs. buy, from the buyer's chair: integration risk, total cost of ownership, and what the roadmap looks like either way.

+ Also inside: scaling this to a 100+ account book →

View the Case

Resilience

A vendor's own deployment cycle — discovery, a scoped POC with success criteria set before it ran, objection handling, a technical close — run against Pawsome's real cross-plant infrastructure risk.

Discovery through deployment, not just a demo — run against real infrastructure risk, not a script.

View the Sales Cycle