Company: BrandLovrsGoals: - Database entity and microservice architecture design- Co-developed a microservice that lists challenges that better fit creators tastes using Pinecone as Vector Database and ChatGPT to create textual description of creators- Writing microservices with a keen eye on coding good practices using Clean Code and SOLID principles- Developed TS-written, internal libraries (field validations, IoC - automatic dependency injection, CLIs, etc.) to enhance development experienceChallenges:- Target business opportunities and bottlenecks to solve by proposing architecture solutions that better fit business rules and needs.- Improving developer experience workflow by writing internal-use, TS-based libraries for different purposes (e.g. field validation, IoC, caching, etc.)- Implemented an important feature to integrate, deliver notifications to creator contents via One Signal, by batch-loading users to be notified to the memory and calling One Signal's API once ready. This allowed notification service to run without topping lambda's memory usage and avoid message loss, since One Signal's API has its limits.Techs:Node.js, TypeScript, JavaScript, MySQL, PostgreSQL, Redis, AWS (Lambda, S3, SQS, API Gateway), Gitlab CI, Pinecone, Golang/Go

Company: FanHeroGoals: - Migrate FanHero API to its 4.0 version, written with Nest.JS/TypeScript/MongoDB from a monolithic core written in Elixir/Ecto/Phoenix stack.- Supported the monothic server while also working in the transition- Documented FanHero whole architecture and design new features having architecture design as key factor in order to better support business decisions- Enhanced internal libraries so they can support new system features- Documented both internal GraphQL client API and developer APIChallenges:- Proposed architecture solutions to solve product team problems, taking DRY principle to the highest level. Enhanced internal libraries to optimize development experience.- Make sure integrations with all microservices would work flawlessly, since most of the features were able to be called both as services (as MVC service layer) and as a RPC functionTechs:Node.js, TypeScript, Elixir, Phoenix, Ecto, MongoDB, Mongoose, Apollo GraphQL Federation, ReST, Swagger, Redis caching, and microservice communication over RPC with Redis, Nest.JS, CASL

Company: Compass UOL / Grupo DPSPGoals: - Use hexagonal architecture concepts to design and implemented features for Medication Benefits Program - integration system, using techs like Express.js, Redis, MongoDB and Apache Kafka- Use Jenkins CI and GitLab to release tagged code versions. Used code versioning good practices to enhance error trackability- Implemented an admin/backoffice system to control benefit authorizers and query benefit program customersTechs:Node.js, TypeScript, MongoDB, AdonisJS, Kibana, Jenkins CI, GitLab, Git, Express.js, GraphQL, Docker, Swagger, Redis, Apache Kafka,TDD, Dynatrace, Jest, Hexagonal Architecture, SOLID

Company: AbacommGoals: - Implemented an AWS Lambda-based serverless backend for a notary office internal system, using DynamoDB, AWS Lambda, AWS S3 static website hosting, Node.js/JavaScript- Implemented a GCP App Engine hosted, AdonisJS-based, web server for PwC Balanced Scorecard system. Also implemented the data flow integration between its Vue.js-based client and the web server with a keen eye to accessibility. DEV environment would still run in an AWS environment.Challenges:As a contractor I needed to migrate the system to a new version with massive breaking changes, but maintaing the old BSC running. We opted to have a LB (Load Balancer) to switch in-out systems versions and make switching systems (which would happen very often) a smooth, low-friction experience.Techs:Node.js, PM2, AdonisJS, MySQL, Vue.js, Vuex, JavaScript, Nuxt.js, GCP bucket static website hosting, GCP App engine, GCP SQL, GCP LB (Load Balancer)

Company: SEDEPTI-BC UFPAGoals: Migrated FICAT, a system for automatically generating catalog cards and has PDF card generation modules, statistics and graphs on card generation (with Chart.js). It was developed in a Node.js environment (with Koa .js) and Vue.js frontend, embedded in a project with Nuxt.js, focusing on user accessibility good practices.Challenges:Target security issues and treat them implementing anti-XSS and CSRF features, implemented request XSRF tokens and Used Knex.js to avoid SQL injections and optimize queries when executing queries for Reports moduleTechs:Node.js, Koa.js, JavaScript, Knex.js, Bookshelf.js, SQLite, Nuxt.js, Vue.js, Chart.js, Jest, TDD, SOLID

Company: LabES (Software Engineering Laboratory)Goals: - Propose and implement WebAPSEE software artifact upload, listing and download to Dropbox and Google Drive (rather than old CVS)- Migrated Software Process Marketplace (SPM) to a new version, using SpringBoot/Nuxt.js stack. The process editor was made under mxGraph (same library as https://diagrams.net, former draw.io) and made collaborative with Apache Kafka, so when some process is being edited it will broadcast changes to all clients listening at the process, and of course, save all the changes based on the mxGraph-generated graph XML.Techs:SpringBoot, JHipster, JPA, Hibernate, MySQL, Java 8, Nuxt.js, Vue.js, JavaScript, Buefy, Apache Kafka