Michael Martiradonna

I move customers and their data onto new systems, and I build the tooling that proves it landed.

Implementation & Solutions Engineer. Data migration, integrations, Python/SQL. AWS Certified Cloud Practitioner. Remote, Chicago area.

Five years of customer-facing IT, from a federal medical center's helpdesk to migrating a fitness business onto a new CRM and course platform. Three shipped projects below: a fail-closed migration validator, an ML service with CI and a hosted demo, and a rules-driven coaching app.

Projects

Three, each with a live link, the code, and the proof

Headline project · data migration

Cutover

A fail-closed toolkit for moving a legacy export into a new system: validate against reviewed rules, map columns, reconcile source against target field by field, upload to S3 with hash verification, and sign off with a runbook. It repairs nothing and blocks the handoff on any failure.

34 automated tests · two worked examples (a property-management tenant export, a legacy parts list into a cloud table) · independent code review · byte-for-byte field comparison and SHA-256 checksums

Python · YAML · SHA-256 · boto3/S3 · GitHub Actions

Code Video (soon)

PS> python validate.py tenants_appfolio_200.csv --rules rules/tenants.yaml --out run-01
Exit code: 1
row,rule,why
17,R001,unit is required; it is blank.
42,D001,move_in_date must be an exact YYYY-MM-DD date.
88,E001,"status must be exactly active, notice, or vacant."
131,U001,Duplicate unit and name; also in CSV record(s): 132.
… 12 rejections. No clean.csv written. No checksum. Human review required.

PS> python reconcile.py run-02/clean.csv run-02/target.csv --key unit --out run-02
Verdict: PASS (14 checks, 0 failed)

ML service, end to end

Ames Home-Price Lab

Describe a home, review the extracted details, and a saved Random Forest estimates its historical sale price in Ames, Iowa. Training-fitted preprocessing, five tracked experiments, validation-based selection, an exported pipeline with a checksum manifest, Docker, and CI that runs tests, training and a container check.

MAE $17,765 on 586 held-out homes, 72% below the median baseline · 83 tests · graded 91/100 · hosted demo runs with no API keys

scikit-learn · MLflow · DVC · Streamlit · Docker · GitHub Actions

Live demo Code Video · 5:45

Ames Home-Price Lab showing a historical estimate of $219,776 and its explanation

Rules-driven coaching app

Gym Swap

When a planned exercise's equipment is busy or missing, Gym Swap offers a reviewed substitute and explains the tradeoffs. A deterministic rules engine decides; an optional explanation layer only cites reviewed coaching notes, so no model ever invents advice. The first public piece of seeform, a longer project on movement coaching.

133 exercises · 270 reviewed comparisons · 357 automated tests in the source suite · shipped through a Lab-first release process with phone acceptance checks

JavaScript · esbuild · retrieval over reviewed notes · Cloudflare Pages

Live demo Code Video (soon)

Gym Swap on a phone showing a reviewed swap card

About

Why implementation work

I spent a year at the Captain James A. Lovell Federal Health Care Center resolving 100 to 150 tickets a week for clinical staff with 100% satisfaction, and before that I ran the technology for a fitness business: a CRM, a mobile app and a course platform, including migrating every client's records onto them without losing history. The pattern in both jobs is the same one implementation teams hire for: understand what the customer actually has, move it correctly, and make sure they can use it.

I also build. The three projects above are the evidence: a migration validator that refuses to guess, an ML service with honest metrics and CI, and a coaching app with 350+ tests. Outside work I am a NASM-certified trainer and run Chi-Town Performance Lab; seeform is where the coaching and the engineering meet.

AWS Certified Cloud PractitionerTripleTen Data Analyst + AI/MLRoosevelt University, 130+ creditsNASM CPT · PES · CESPreviously held TS (2013) and Secret (2023)

Experience

Customer-facing technical work since 2017

12/2023 – 12/2024

IT Helpdesk Technician · Captain James A. Lovell Federal Health Care Center (VA)100–150 tickets a week, 100% satisfaction, two Bravo Zulu commendations. Active Directory and ServiceNow under federal change control.

2021 – 2023

IT Technical Manager & Sales Support · Private Lending GroupText-based lead-qualification automation saving 10+ hours a week; directed contractors on web and CRM work.

08/2022 – 12/2022

Junior Infrastructure Apprentice · Cloud5 CommunicationsDevice fleet updates and remote resets for major hotel chains; automated error reporting.

2017 – 2020

IT Strategist & Technician · Elite FitnessLed the migration of client records, scheduling and course content onto a new CRM, app and education platform; Zapier, Manychat and ActiveCampaign automation saving 80+ hours a month.