David Martin Garcia

I joined the company as part of the Returnly acquisition and quickly took charge of the Comms Platform team, which manages the infrastructure for sending and receiving user communications.

Our existing codebase was shared across multiple teams and we started setting the ownership of the team on this service, acting as providers for other teams and products.

We started performing refactors and rearchitecturing the code as part of an effort for supporting future spikes in trafic, we recognized these challenges and took decisive action, ending in their first successful BFCM year without any incident.

We developed internal tools that enabled us to send emails to +1M users efficiently. I spearheaded the upgrade to Python 3, tackling incompatibilities with unmaintained libraries head-on.

I conducted load testing to determine necessary resources for Black Friday/Cyber Monday (BFCM) and critical launch dates for our merchants.

I provisioned resources using Terraform and Terragrunt, ensuring we met regulatory changes and security protocol updates, including DMARC and SMS regulations.

I also tracked user communication events and streamlined this data into the analytics team, played a key role in our UK launch, and proactively traced and resolved performance issues within the test suite.

I joined the Platform team, where I have accomplished the following:

- Upgraded Ruby from version 2.3 to 3.2, which is the latest available version.

- Upgraded Rails from version 4.2 to 7, the most recent version.

- Reviewed and improved access to the database, using read-only replicas.

- Managed the compatibility of gems and libraries with the new version upgrades, which involved collaborating with, contributing to, or forking several gems.

- Perform planning for continuous code improvement, handling the reshaping of the architecture.

- Addressed issues in the continuous integration (CI) process by speeding it up and removing flaky tests, thereby enhancing overall productivity across the company.

- Improved Docker builds to make them smaller and set up a multi-build CI to support the version upgrades.

- Implemented a search engine with Elasticsearch for showing the user suggestions about products and merchants.

I joined Sonnen as part of a new project to revamp the overall architecture. I was part of the Platform team, which managed the infrastructure necessary for handling the collection of the data from their main product, solar battery assets.

Initially, we relied on a third-party provider for our database to store metrics. Unfortunately, this solution was unable to scale effectively, and costs were higher than expected by the leadership.

We set a project to migrate the event collections system to our own database, using multiple shards and partitions to improve performance and availability. We improved the code and queries to better handle the load, which resulted in cost savings and a more resilient product compared to the third-party solution.

We also dockerized the entire environment moving away from Capistrano to a Docker-based deployment. As the architecture and the organization evolved, we adapted to serve and create REST APIs for various new teams and initiatives. We exposed these APIs to the public and the company’s applications, implementing an Authorization, Authentication, and Permissions layer for all the underlying microservices.

We successfully managed a 15 TB database that operated flawlessly, compacting metrics and ensuring consistent performance, with a caching layer to collect statistics, eliminating the need for continuous recalculation.

We created a payslip calculator engine for multiple regions (ES and UK), automatizing the process at the end of the month of creating the bank files and having all the PDFs for multiple companies.

I also managed to provision and maintain resources on AWS. Additionally, I transitioned our infrastructure from Fargate and EC2 to ECS and established infrastructure monitoring.

Lastly, I improved the synchronization process across multiple integrations with 3rd parties, making them faster and more optimal, not querying the same data multiple times and setting the architecture based on Sidekiq and a custom state event machine.

We started as a two-person team and developed an application for HR departments that collects and aggregates information on job listings. Our first successful product was an applicant tracking system that gathered candidates from various sources.

To enhance our platform, we implemented a search engine based on Elasticsearch, that allows recruiters to filter profiles based on specific keywords and questions. This involved handling client requirements and negotiating upcoming features.

We introduced a bot-assisted job publishing and collection engine, incorporating techniques to bypass bot detection algorithms across 40+ job portals. This system remains effective today.

To ensure efficient data collection, we improved database performance, enabling our bots to gather thousands of job offers daily. We also added multi-tenancy for specific clients and developed white-label web pages for them.

Additionally, we created a Chrome extension that allows recruiters to send CVs and profiles from major business social networks and job portals. We built a CV parser using text mining techniques.

Finally, we migrated from bare-metal servers to AWS, provisioning our infrastructure using Terraform and Ansible.