David Wuscovi
MenuClose

Cloud Solutions Architect · Tech Lead

AWS architecture and technical leadership with results.

I design platforms on AWS, get systems ready for the most demanding traffic peaks of the year and organize how teams work so every production deployment is safe. I've done it in retail and in fintech.

Consulting Contact

Traffic peaks
Stress testing and Redis caching for Hot Sale, CyberMonday and Black Friday
Cloud
Payments 100% on AWS, with microservices on Kubernetes (EKS)
Teams
Several full-stack developer squads led in parallel
Integrations
MODO, Banco BIND, Banco Coinag and Newpay, with payment reconciliation
Reference architecture: users, API Gateway with Cognito, services on Kubernetes (EKS), events on Lambda and the business outcome, with data on RDS and files on S3.Userstraffic peaksAPI GatewayCognito · rate limitServicesKubernetes · EKSEventsasync · LambdaBusinessorder · payment · noticeDataRDS · PostgreSQLFilesS3Reference architecture for high traffic on AWS.Userstraffic peaksAPI GatewayCognito · rate limitServicesKubernetes · EKSEventsasync · LambdaBusinessorder · payment · noticeDataRDSFilesS3
Reference architecture for high traffic on AWS: the foundation for both an e-commerce site on Hot Sale day and a QR payment.

Experience

Results in production.

Two real examples of what I do: what needed solving, how I solved it and with what technology.

MEGATONE

Leading home appliance and technology retailer in Argentina · more than 60 stores in 20 provinces

Technical lead for the chain's e-commerce, including its Hot Sale, CyberMonday and Black Friday.

  • I was in charge of the site for every Hot Sale (Argentina's biggest online sales event), CyberMonday and Black Friday: preparing the platform, deploying and running it through the year's traffic peaks.
  • I led several squads of 3 or 4 full-stack developers focused on e-commerce, more than ten people in total.
  • I designed and ran stress tests to find the platform's limits before each event.
  • I implemented Redis caching to sustain the load at the highest-traffic moments.
  • I integrated MODO as a payment method for the e-commerce site.
See 4 more results
  • I worked with SEO agencies and we built a system that generates static pages for key sections of the site, which improved its search visibility.
  • I built the electronic invoice printing system for the more than 60 stores across the country: each invoice prints on its own store's laser printer.
  • I was responsible for production deployments on the company's own on-premise servers.
  • I represented MEGATONE at the Google Tagathon: Modern Web Tracking, a Google workshop on measurement with Tag Manager, dataLayer and Google Analytics, alongside digital teams from large companies.

C# · ASP.NET Core Blazor · Vue.js · SQL Server · Redis · APIs REST · Google Tag Manager

BenkoPay

Fintech pioneer in QR payments since 2017 · registered with the BCRA (Argentina's central bank) as a PSP and PSPCP · current

I designed and built a payments platform 100% on AWS for interoperable QR payments and CVU collections.

  • I led API QR and API CVU, the company's two core products: interoperable QR payments within the BCRA's regulatory framework and collections through CVU (Argentina's virtual account numbers) for invoices and recurring payments.
  • I led the tax engine and the commission engine, two core solutions the company had needed for years.
  • I migrated the platform from a Docker Compose monorepo to microservices on Kubernetes on Amazon EKS.
  • I integrated the platform with Banco BIND, Banco Coinag and Newpay.
  • I implemented payment reconciliation, so every charge matches what the banks credit.
See 4 more results
  • I solved user authentication with Amazon Cognito.
  • I built the CI/CD pipelines in Bitbucket and I'm responsible for production deployments.
  • I implemented Grafana to monitor the platform in production.
  • I lead the development team with agile methods and project management in ClickUp.

Node.js · NestJS · Next.js · React · AWS EKS · Lambda · Cognito · Grafana · Bitbucket Pipelines

Education
Higher Technician in Data Science and Artificial Intelligence (ISTEA) · DevOps Engineer (EducaciónIT)
Based in
Argentina · remote work with teams worldwide

How I work

Six principles I apply on every project.

Prepare the platform before the peak

Capacity gets tested and tuned before the event, not during it: stress tests to find the limits and Redis caching to raise them. That's how I handled every Hot Sale, CyberMonday and Black Friday.

Automate the path to production

CI/CD pipelines from day one, so every deployment is repeatable. I was responsible for production deployments in both retail and fintech.

Migrate without stopping operations

The legacy system stays in production while the new one is built alongside it, and traffic moves over in stages. That's how I moved a payments platform from a Docker Compose monorepo to microservices on Kubernetes without interrupting payments, and migrated parts of a .NET e-commerce site while it kept selling.

Auditable money flows

Every banking integration includes payment reconciliation, so every transaction can be traced end to end. I applied it with MODO, Banco BIND, Banco Coinag and Newpay.

Architecture the team understands

Services with clear boundaries and architecture decisions in writing, so the team moves forward without depending on a single person.

Technical judgment in every review

I read every change before it reaches production, whether a person or an AI agent wrote it. I suggest and teach with concrete examples, something I also did as a university instructor.

Applied AI

I direct AI agents and review every result.

Today AI agents write most of the code. My work comes before and after: defining the architecture and the criteria, and reviewing every implementation before it reaches production.

  1. 01

    Define

    Architecture, service boundaries and acceptance criteria, in writing.

  2. 02

    Delegate

    Well-scoped tasks to AI agents, with the context and constraints they need.

  3. 03

    Review

    Design, security, tests and cloud cost of every change. What doesn't meet the bar goes back.

  4. 04

    Deploy

    Only reviewed work goes through the CI/CD pipeline to production.

deitafix

An AI agent can propose writes to a production database through MCP, or turn a plain-language request into SQL, but a person always approves the execution.

This site

I built it by directing an AI agent: written criteria for content, tone and design, and a review of every change before publishing it.

Education

Higher Technician in Data Science and Artificial Intelligence (ISTEA).

AI speeds up the work. Responsibility for what reaches production is still mine.

More

Consulting, open source, training and blog.

ConsultingAWS architecture, legacy system migration, payment systems and technical leadership, up to twelve hours a week.
Open sourcedeitafix: safe writes to production databases, with AI agents that propose and a person who approves.
Teaching and trainingI trained Android developers for Industry 4.0 and took part in the Google Tagathon at Google Argentina.
BlogShort notes on outages, migrations, payments and AI in production, each with its source.

Contact

Let's talk.

Tell me about the context and what you need. I reply personally.

Topic