Hello! On November 30, in our office, together with the Golang Moscow community, we held a meetup on the occasion of the decade of Go. The meeting discussed machine learning in Go services, multi-cluster balancing solutions, techniques for writing Go applications under Cloud Native, and the history of Go.

Go under the cut if you are interested in these topics. Inside the post - all the materials of the meeting: video recordings of reports, presentations of speakers, feedback from meetup guests and links to a photo report.

Reports

10 years of Go — Alexey Palazhchenko

A report about the past and future of Go, its ecosystem and its communities, including Golang Moscow.

→ Presentation

Listener reviews

• Learned a lot from the history of Go. It was interesting.
• It was interesting to learn about the history of the development of the language and the community.
• There would be more such people and reports!

Integration of ML models into a Go service — Dmitry Zenin, Ozon

The story of how Ozon applied machine learning to category prediction. The experiments were carried out using python and its ecosystem for ml. However, production in the company lives on go and Dmitry talked about how they implemented their developments into an existing go-service, what metrics they covered it with and what they got as a result, both in terms of the initial task and in terms of the performance of the entire system .

→ Presentation

Listener reviews

• The report is not for everyone. It will be of interest to those who are fond of ML, neural networks and so on.
• Case from real development. It's always nice to hear about implementation from idea to implementation.
• At my last job, my initiative was to move the generation of variables for machine learning models to Go. It went into prod. It was interesting to hear how people connected Tensorflow/fasttext.

Navigator - solution for multicluster balancing: reasons for creation and development experience - Mikhail Shaverdo, Avito

Mikhail spoke about the features of developing and testing cloud-native applications on Go using the service mesh in Avito as an example.

The program includes:

• why Navigator is needed: several DCs and Canary;
• why third-party solutions are not suitable;
• how the Navigator works;
• unit tests are good, but better with e2e;
• pitfalls we have encountered.
  

→ Presentation

Listener reviews

• Interesting, but I'm not a devops. I recommended it to a friend, he will be interested. Especially since he, too, began to encounter canary releases.
• There were many new things for me. I could not understand everything, but the performance was still interesting.
• I am learning Kubernetes. The report is very helpful.

Preparing services for the world of cloud infrastructures — Elena Grahovats, N26

Go is one of those programming languages ​​that you fall in love with seriously and for a long time. However, to start writing in it effectively, it is not enough to learn the syntax and go through the Go Tour or read the textbook. Elena told what tricks will be required to write Go-applications under Cloud Native, how to work with external dependencies as safely as possible, and how to correctly dockerize services written in Go.

→ Presentation

Listener reviews

• Super report. Very useful and directly applicable in practice.
• Tells interesting. Lots of interesting cases. In general, the performance was positive.
• Good advice. Maximum practice.

references

Playlist with all the videos from the meetup can be found on our YouTube channel. In order not to miss the next meetup in Avito, subscribe to our page in Timepad.

We posted photos from the meeting on the AvitoTech pages in Facebook и VK. Take a look if you're interested.

Until next time!

Source: habr.com