Gonzalo Martin PerezAI Software Engineer

Work

Experience and selected projects

2 years and 8 months of professional engineering, as of September 2026. Product-scale figures describe the systems; what I did is listed separately underneath.

Professional experience

  1. Sep 2026 – Present

    Argentina · Remote

    AI EngineerRampy

    AI engineering in a fintech context.

    • Set up an agentic engineering workspace that preserves context and enables coordinated parallel work.

    Recently started. Only work that is finished and cleared for publication appears here.

    • Applied AI
    • Agentic systems
  2. Dec 2025 – Aug 2026

    Argentina · Remote

    Software EngineerTeamcubation

    The promotions platform for Payway, a leading Latin American fintech, serving an ecosystem of 350,000+ merchants, 70+ card issuers and 13 countries.

    • Delivered the promotions experience end to end: a Single-SPA React and TypeScript microfrontend embedded as a native section of the merchant portal, connected through a Spring WebFlux BFF to multiple microservices and providers.
    • Used Amazon API Gateway to bridge the HTTPS merchant portal with internal HTTP enterprise services, keeping the integration secure from the application side.
    • Built Java/Spring Boot and TypeScript/Node.js/NestJS microservices to ingest, normalise and process promotions from merchants, financial entities and providers, adapting contracts per integration.
    • Implemented an event-driven, serverless flow with Amazon S3, SQS and a Python/FastAPI Lambda using ordered processing strategies.
    • Designed an in-product RAG and agentic assistant for merchant owners with Python, LangChain, LangGraph, PostgreSQL/pgvector and guardrails, retrieving the banking promotions tied to each merchant to ground the model's answers.
    • Built Python and Node.js tooling to replay bulk loads safely against isolated local databases, reconcile results and surface failed promotions and edge cases before release.
    • 20M+

      promotions loaded

      The complete historical load, executed in production, followed by thousands of new promotions daily.

    • ~90,000/s

      peak throughput

      Measured during production execution with real data across the full 20M+ load.

    • Zero

      data loss

      Validated across that same complete production load.

    A key engineering team member with substantial ownership of these components, not their sole author. Worked on containerised services in Docker and Kubernetes environments as a developer — configuration, observability and diagnosis — not as cluster administrator or infrastructure owner.

    • Java
    • Spring Boot
    • Spring WebFlux
    • TypeScript
    • Node.js
    • NestJS
    • Python
    • FastAPI
    • React
    • Single-SPA
    • PostgreSQL
    • pgvector
    • LangChain
    • LangGraph
    • AWS Lambda
    • Amazon API Gateway
    • Amazon S3
    • Amazon SQS
    • Docker
    • Kubernetes
    • OpenTelemetry
    • GitLab CI/CD
  3. Jan 2024 – Present

    Argentina · Remote

    AI Engineer & Full-Stack DeveloperSelf-employed

    Independent product engineering for e-commerce, social media and digital-product clients. Projects are occasional rather than continuous.

    • Designed and shipped digital products, back offices and stock and point-of-sale extensions, connecting commercial needs to maintainable systems.
    • Built MCP-, Claude- and LLM-powered pipelines and agents that research, generate and publish content end to end.
    • Developed online stores, WordPress solutions, APIs and custom full-stack integrations, with Python and Node.js tooling for automation, bulk processing and data analysis under spec-driven development.
    • USD 10,000+

      cumulative client profit

      Profit generated for clients by the products delivered — not personal revenue.

    • ≥70%

      operating-cost reduction

      Estimated by the clients themselves. Not an audited financial measurement.

    Client count and client names are not published, because that history was never reconstructed precisely enough to state.

    • Python
    • Node.js
    • TypeScript
    • Claude API
    • Model Context Protocol
    • REST APIs
    • WordPress
  4. Dec 2024 – Dec 2025

    Bahía Blanca, Argentina · On-site

    Software EngineerCooperativa Obrera

    SGA, a secure enterprise back office for administering sensitive data, permissions and auditing at Argentina's largest consumer cooperative: 150+ branches, 2.7M+ members and an ecosystem of 50+ internal systems.

    • Held a principal design, architecture and implementation role with product-engineering scope.
    • Built a responsive Next.js and TypeScript frontend whose only interface was a Python/FastAPI BFF, which defined the shared contract and decoupled the web experience from every downstream service.
    • Personally integrated the 10+ systems within SGA's scope through Java/Spring Boot or TypeScript/Node.js/NestJS microservices, each conforming to that contract and encapsulating its own internal logic; legacy PHP applications were brought in through dedicated integrations.
    • Standardised onboarding for new systems with reusable structures and filters.
    • Automated bulk loads and data validation with Python, and implemented LDAP authentication and a complete audit trail with exportable reports.
    • 10+

      systems integrated

      Every system within SGA's scope, integrated personally, inside an ecosystem of 50+.

    • ≥80%

      fewer permission tickets

      Compared against tickets and logs before and after standardised onboarding; ≥95% of recurring assignment bugs eliminated.

    • 10 min → <10 s

      permission search

      Through SQL tuning, indexing and pagination. Critical endpoints stayed below 300 ms on average in qualified, repeatable tests on on-premise enterprise servers.

    No public links exist: SGA administers back offices, personal data and internal company processes.

    • Python
    • FastAPI
    • Next.js
    • React
    • TypeScript
    • Java
    • Spring Boot
    • Node.js
    • NestJS
    • SQL
    • Docker
    • LDAP
    • Prometheus
    • Grafana
    • GitLab CI/CD

Selected projects

Work with public evidence

One project is published in full, with its source. The systems I built at Cooperativa Obrera and Teamcubation administer internal company data, so they have no public links — the experience above is the record, and I'm happy to walk through either in a conversation.

Filomena

A health-sciences exam platform five institutions run on

Jan – Dec 2025

A legacy CakePHP and jQuery monolith, rebuilt as an API-first platform that now runs high-stakes exams for five Argentine national institutions.

Built by a three-person team. I was the principal author and contributor across architecture, backend, frontend, infrastructure, security, observability and coordination. The platform is a team achievement.

  • 5

    institutions in production

    UNS, UNRN, UNC, UNVM and FAMFyG.

  • 1,000+

    simultaneous users

    Observed in production with complete data consistency, not a synthetic benchmark.

  • <300 ms

    critical endpoints

    Average latency, through SQL tuning, Redis caching and queues.

  • 10 / 10

    final year project

    Graded at Universidad Nacional del Sur.