GUILLERMO CODURI

As the Tech Head (Manager) of the Patterns Services area (15+ developers), our focus is on building capabilities on an in-house platform that abstracts from well-known cloud services such as AWS and GCP.

Within this area, we've developed services designed to be easily, efficiently, and resiliently utilized by Mercado Libre developers. The core idea behind Patterns Services is to avoid constantly reinventing solutions and instead offer recognized service patterns to our clients (developers).

Among the services we've built are:

• Streaming data - Using Kinesis Stream.
• Event Sourcing - Using Dynamo DB with Kinesis Streams.
• Messaging Queue - Leveraging Google Pub/Sub.
• NoCode Processing - An in-house developed product similar to AWS Pipes & Bridge.

All these services are offered generically to developers, abstracting away the specifics of cloud providers

Rotation to Platform Services as Team Lead.

In this area, our objective is to build products and infrastructure for development teams, aiming to enhance productivity, quality, reliability, stability, and delivery time. Additionally, we collaborate to cultivate an exceptional developer experience and improve developer engagement.

On a daily basis, we:

• Create and manage the release pipeline, including architecture, infrastructure, and development.
• Develop and maintain dev and production Docker images.
• Implement safe deployment tools and release scripting.
• Design and implement testing tools, frameworks, and distributed load test solutions.
• Provide cloud computing solutions and high-load log management tools.
• Automate analysis of repository, configuration, and application health status, ensuring adherence to best practices.
• Monitor applications and orchestrate recovery processes.
• Collaborate closely with numerous teams utilizing various programming languages and technology stacks such as Java, Go, Python, Node.js, etc.
• Analyze data using Big Data tools.

As a Tech Team Lead Engineer, main responsibilities was:

• Developing and maintaining services (building blocks) for the entire IT area, serving approximately 4,000 developers, in Platform-as-a-Service (PaaS) mode.
• Focusing on communication services employing a pub/sub model, as well as streaming and processing large volumes of data, with a strong emphasis on high availability and performance.
• Leveraging cloud platforms and technologies including AWS, GCP, Pub/Sub, Kinesis Streams, Kinesis Analytics, DynamoDB, CloudWatch, and Route53, alongside programming languages such as Spring Boot, Python, Go, Ruby, React, Quarkus, and Spring WebFlux.
• Leading a Technical Team of 8 Developers, providing mentorship and fostering continuous improvement while maintaining a positive working environment.
• Conducting technical interviews as part of the hiring process.

Developed and led a team as a Sr Engineer to preventing fraud in the marketplace, resulting in preventing $300k USD monthly.

Main Responsibilities:

• Designed, implemented, and maintained architecture components within the machine learning pipeline.
• Coordinated with business analysts to evaluate various fraud cases and devise preventive measures.
• Defined and measured key performance indicators (KPIs) for projects.
• Provided regular updates and roadmap presentations to stakeholders.

Technologies Used:

• Spring Boot: Java framework for backend components.
• ElasticSearch: Used to serve features of online predictive models.
• Python + Parquet + Apache Spark: Orchestrated execution of machine learning model training pipeline.
• Scikit-Learn + Keras: Main training libraries utilized.
• AWS Neptune: Employed as a graph database for link analysis.

Additional Responsibilities:

• Conducted technical interviews as part of the hiring process.

Work on innovative solutions within the fraud prevention team, empowering users in the platform to self-manage their situations upon detection of fraud risks in their behavior.

Led the analysis, design, and implementation of diverse identity verification (Know Your Customer - KYC) mechanisms across multiple platforms including Web, Android, and iOS. These mechanisms encompassed identity document scanning, cell phone validations, selfie-based proof of life challenges, credit/debit card validations, and user recommendation systems within the Mercado Libre platform.

Use different technologies and frameworks to ensure robust and efficient system architecture, including:

• Spring Boot framework, Java 11.
• React JS for frontend.
• Objective-C for iOS native app modules.
• MongoDB for storage and management of user data and metadata.
• Leveraged Convolutional Neural Networks (CNN) for accurate detection of objects in images, enhancing fraud detection capabilities.
• Integrated with Google Cloud for Optical Character Recognition (OCR) detections, enabling seamless extraction of text from images (identity documents).
• Collaborated with FaceTech for 3D face validations, enhancing the security and reliability of identity verification processes.

Designed and implemented an Tax algorithm for the Uruguayan Government, facilitating the comprehensive calculation of taxes for citizens with a historical perspective.

Orchestrated the integration of the algorithm into a microservices architecture, enhancing scalability and maintainability.

Developed with REST APIs to provide seamless access to tax calculation functionalities, ensuring ease of use and interoperability with other systems.

Worked with Java Spark and MySQL.

Develop a public website for the Uruguayan Government, enabling citizens to conveniently submit forms for online paperwork with electronic digital signatures.

Implemented the website using JSP Pages, ensuring a user-friendly interface and seamless navigation.

Configured and maintained the website on a Tomcat server, ensuring optimal performance and reliability.

Established a backend using MySQL server to store and manage submitted forms and electronic signatures.