Ivan Porto Carrero 🦇🔊

St Gummarus college

San Francisco Bay Area

Netflix

Senior Software Engineer

I'm the architect at One Concern. This means I'm responsible for providing technical oversight for the entire system.

In this case this means introducing CI/CD, kubernetes, serverless and datascience pipeline governance of data and models.
I also mentor the engineering team members, work with product on definitions and help define the process for the company to scale and keep up with the rapid growth

During my time here I've grown the team from 3 engineers to 15. Migrated the application from ruby on rails managed with a chef deployment without any meaningful CI process to a multi-environment CI/CD process that deploys on kubernetes.

When I started here the company had 2 different VPN's, authentication and 2FA was done by different services, I consolidated those into a beyond corp security model with Google as IDP for all the internal services.

Migrated from rails' devise auth system to using Keycloak because of SSO needs.
Replaced the usage of homegrown CSS/JS framework to React for more predictable results.
Replaced the remaining functionality of rails to a go based microservice architecture where the public facing API is graphQL but the internal services use GRPC.

Integrated end to end encryption both at rest and in-transit on top of the kubernetes deployment.

Introduced argo to the machine learning team so they can build large workflows by composing different containers. And moved most of the ML data processing to be filesystem based instead of postgres.

One Concern

Architect

I started a member of the Open Source Team in a nascent BU called CNA (Cloud-Native Applications).

This role is a bit like an internal Open Source Evangelist. I took it to mean popularizing modern approaches to API development. It also means that I function as advisor when it comes to integrating key technologies with cloud-native applications like Mesos, Kubernetes, docker,...

During this period I got to work on go-swagger (a swagger implementation and toolkit written in go) which is now being used by several teams to build and document API’s. Currently I’m supporting those
teams by fixing bugs in swagger, provide code reviews.

After a few months I was put on a project to automate running self-healing stateful services (databases, message queues,...) without user intervention. We needed to focus on just kubernetes, and I became tech lead on integrating kubernetes on vSphere. This role makes me cooperate with many different business units in the company and provide the direction as well as design of making kubernetes a first class citizen on vmware’s flagship product vSphere. This project was ready for shipment when we were
reassigned to work on a project named PKS.

I was moved as overall tech lead for VMware on the PKS (Pivotal Container Service) project. This project is also running kubernetes on vsphere, but this time it uses pivotals lifecycle manager called BoSH. My role is to provide direction and ensure the team of 40 engineers is successful for integrating things like
NSX-T, the vRealize suite and wavefront.

Non Project work:
-  Collaborated with VRG (VMware Research Group) on a better approach for gossip called Rapid.
-  Socialized swagger/openapi 2.0 through a series of tech talks and meetings with the VMware
API team
-  Prototyped using ESXi as a native target for a kubernetes container runtime (a CRI
implementation for esxi). This is now being used by 2 distinct teams to base their approach on.

VMware

Staff Engineer

The current technology stack in use at reverb is: MongoDB, Redis, Scala (Scalatra, JSON4S, Akka, ….), Java, Node.js, Go, AWS, ElasticSearch and RabbitMQ. There is some hadoop involved as well.

As the only architect at Reverb, I designed the system that powers the Reverb iOS application. I replaced an architecture that saw backend requests ranging from 2 or more seconds (up to 90 seconds in some cases) with one where the response times are generally below 1s and mostly around 300ms.

In addition to designing and leading the development of the backend infrastructure, I:

-  recruited a couple of excellent engineers to help us build the iOS backend infrastructure.
-  worked with data science in an advisory role, focusing on designing new systems for data science
-  introduced continuous delivery
-  transformed to frictionless deployment process with Docker on Ubuntu (replacing RPM/Centos), enabling resource sharing & leading to huge cost savings
-  built a minimal framework for the JVM applications to hook into service discovery and the like
-  decommissioned a homegrown cumbersome service discovery and replaced with a Zookeeper-based one
-  introduced logs and system metrics as event stream
-  implemented Datadog for dealing with the stream of metrics events and interesting log events, replacing Nagios and the sound-of-ignorance
-  led development team in optimizing based on newly-available metrics
-  redesigned the machine learning infrastructure to leverage elasticsearch rather than a very large homegrown system that was becoming outgrown

Go-Swagger lead, Scalatra lead, JSON4S maintainer.

Reverb Technologies Inc. FKA Wordnik

Senior Architect

Adam Burmister and I got accepted into the O2 incubation pilot programme.
I can't detail the journey this project took here as its pretty lengthy and requires the entire story so that nothing can be taken out of context. So feel free to ask me about it in a face to face meeting.

Output of this project:
-  Scalatra web framework (in use by BBC, gov.uk, IGN, Netflix, ...)
-  hookup for reliable websocket messaging (fault-tolerant etc, alterntive to akka's remoting transport which suffered from teething problems)
-  Forked lift-json into what would become Json4s

Additionally during this time:
-  spoke at scalalol about leveraging mongodb with scala
-  spoke at the Guardian about our project and how we leverage the JVM ecosystem
-  got awarded Microsoft MVP for C#/IronRuby

BackChat.io

Founder/CTO

Consulting while building a product for the cloud, this continued into the next job where we got into the incubation programme for O2 Telefonica. Please refer to that section for a more detailed explanation.

Additionally during this time I:

-  coached and trained a team to start using ruby on rails instead of <a href="http://ASP.NET">ASP.NET</a> MVC at Selor as a contractor. They were keen to compare it with their existing approach.
-  built a POC collaborative editor backed by XMPP as messaging infrastructure
-  wrote most of the IronRuby book
-  spoke at FOSDEM on IronRuby and Mono
-  Became a C#/IronRuby MVP for Microsoft

White Rabbit Consulting

Owner

At SD Worx there are a lot of developers (±300) and for their front end projects they generally use .NET.  Our team was responsible for creating a framework/software factory in WPF for those developers.

We were also the first team at SD to fully embrace SCRUM and do our work in that way.  In the period I worked there some of the feats I did were: Replace CAB with Prism (only the regions though).  Replace usage of ObjectBuilder with Autofac for dependency injection.

As a member of the architecture team I took part in the discussions of several departments around architecting their web applications etc.

In addition to the architecture work we were also responsible for providing guidance and introduction sessions to best practices etc.  So I also went and advised other departments on how they can solve their current problem with their current technology stack.

I initially wrote their training documentation around BDD with .NET, and provided them with an automocking container based on autofac and mstest.

SD Worx

.NET Solution Architect

Designing, maintaining and developing in-house line of business applications

Elia

Senior Developer

Run projects from requirements gathering up to supporting them in production.

Xero

Ran the business side of things, so everything from looking for customers, on boarding them up to accounting.
Executed most of the projects, outsourcing work I couldn't handle on my own to partners in Russia and China

Flanders International Marketing

See flanders

This company had 3 employees

Koolkraft

Provide support and manage the networks of all satellite offices in Belgium, we are talking about 400 end-users in offices of sizes going from 3 people up to 30 people.