Period: Jul 2025 — present  ·  ● in production

Moises Perez

Sole developer with experience in the full lifecycle from requirements to production.

GitHubLinkedInEmailCV.pdf
Conciliation
Manual workflow
Crossing statements by hand
Inconsistent ERP exports
Ad-hoc AI prompting
Working software
✓ Sella Cruce
✓ Financial module
✓ method-based-harness

I'm an Information Systems Engineering student at UNISON and the only developer at JGRP, building software since July 2025.

I translate non-technical business requirements into working tools — full-stack, with a backend and data lean. The work is mostly listening first, then building the thing that makes the manual version disappear.

Flagship In daily production use

Sella Cruce

Bank-statement reconciliation

It cross-checks movements in bank PDF statements against an Excel of expected transactions and marks the matches — built directly from the by-hand process the team used before.

ProblemStatements were reconciled against expected transactions manually, line by line.ApproachMapped the workflow with the team, then built a tool that parses the PDF, aligns it to the Excel, and surfaces matches and exceptions.OutcomeIn production at JGRP, used daily. Designed, built, deployed and maintained end to end by me.
ReactTypeScriptPythonFastAPIDocker

Internal tool · handles client financial data — no public repo or demo

Reconciliation · sample142 / 150 matched
DateMovementAmountStatus
03/04Transfer ····4821$12,400✓
03/06Deposit ····0937$3,150✓
03/09Fee ····1188$280⚑
03/11Transfer ····5502$8,900✓
Reconciled$24,450✓

illustrative mockup — sanitized, not real client data

Case study · Data / ETL

Financial analysis module

ERP normalization

A Python (pandas) module that normalizes inconsistent trial-balance exports from the CONTPAQi and Aspel accounting ERPs, with data-quality checks that keep results consistent across clients — extraction, parsing, normalization, validation.

Pythonpandasdata quality

Internal tool at JGRP — no public repo

Open source · Early-stage

method-based-harness

Methodology for AI coding

A Python CLI that applies software-engineering methodology to agentic coding harnesses — injecting "methodology packs" into a repo as native Claude Code files. An active exploration of how AI-assisted development should be structured. Runs in Claude Code today; opencode and Copilot support are planned.

PythonCLIClaude Code
View repository →

Schedule of capabilities

What I work with
ALanguages
PythonTypeScriptJavaScript (ES6+)JavaSQL
BFrontend
ReactHTML5CSS3Responsive design
CBackend
FastAPIREST APIsAsync data handling
DData
pandasExtraction / parsingNormalizationData-quality checks
EDatabases
SQLSQLite
FDevOps & tools
DockerGit / GitHubCI/CDn8n
GAI engineering
Claude CodeGitHub CopilotopencodeOllama (experiment)
HAnalysis
UMLRequirements gathering
ISpoken
Spanish (native)English (B1–B2)
Sign-off

Let's build something that lasts.

Open to full-stack, backend / Python roles

Emailmoisesperezaello@gmail.comGitHubgithub.com/moisesPAelloLinkedIn/in/moises-perez
Moisés Pérez
Reconciled & signed
Hermosillo, Sonora
Mexico