Increased demand for analytics, product teams, Amazon, Israel, Singapore, remote work, etc. - discussed a lot.

Habr, hello! Most likely you don’t know, but once a month we meet with our graduates and teachers in the Moscow bar for beer / tea. And in the new reality, new normal, we are forced to bring our traditions into line with reality. We went into the online format with the programs back in mid-March, and last week for the first time we hosted an online meeting of the Newprolab community and were so pleased with the new experience that we decided to write a little about what was discussed and post a meeting on our channel on Youtube(for the convenience of navigating through the video, timecodes were set). We were able to chat with our graduates from Israel, Singapore, a teacher who moved to Scotland. It was nice to see everyone with mugs and glasses, as in a regular meeting at the bar. It was also the first meeting where graduates came with children, and I with a dog :)



Our usual meetings are held without any agenda, but this time we decided to make the meeting thematic, we asked graduates to share what is happening now in their companies / teams, with processes, products, how they change in new circumstances, etc. But in fact they touched upon other topics as at a regular party.

We specifically invited some graduates and teachers to share their situations / news, but they could not join for a number of reasons:

• « , , 12 , . »;
• «, . . . - , — . , , , , , 2-3 »;
• « , , , , . , »;
• « , ».

In this material, we collected some, hopefully, topics that were useful to you that were discussed. Sometimes they will be described very briefly, but you can always see the necessary passage in the recording of the meeting by simply clicking on the desired topic in the description of the video.

According to Nikita Matveev, CDO S7, the company has increased its need for BI reporting, various departments that once had no need for real-time reports now require them. The plus from the situation is that in some departments the business processes stopped or greatly reduced their intensity and now, finally, with them you can implement long-planned projects - as an example, Nikita gives the redistribution of crews.

Then a conversation started about the differences between data science product owner and ordinary product owner, and about product teams in data science. Egor Mateshuk, Quantum, spoke about the difficulty of defining metrics for such teams, because often they do not make the final product, but only part of it. And therefore, it is much more difficult for them to formulate the resulting metrics. The description of the purely model results is not optimal, because they may not be tied to business metrics.

Yegor went on to talk about the state of affairs in their Quantum: their attention to reporting was also sharpened, now they are more focused on sources and monitoring. One of the pains now identified is non-adaptive monitoring. The number of metrics monitored by the company is many, they usually behave calmly, and now due to the decrease in activity there is a series of fluctuating alerts (previously established metrics are not relevant now). Yegor called communication the second challenge. Autonomous product teams work better, and where the functionality is divided between departments, difficulties begin just at the intersections of departments - but there is such a problem “in a normal undeleted life”.

Egor also asked me to share how other companies perceive the creation of a ml-platform for testing models and how this idea is sold by them. Nikita Uchetelev is just developing such a platform at Lamoda, and he told how they build communication with business on this and not only this issue. Sasha Petrov spoke about the SageMaker platform, which they use at Amazon, and which other companies can use as well.

Sasha then spoke about the horizontal organizational structure at Amazon, where product teams are like independent startups distributed around the world and often having duplicate developments. Since the company has always had a lot of video communication with offices in Seattle, Palo Alto, and India, the format is familiar, the difference is that now they meet with their main team not in a conversation, but also online.

Boris Syusyukalov, Delimobil, shared the practice of working with their analytical team in a remote format. If online communication is difficult, you need to minimize it. This implies the existence of an ecosystem that analysts and data engineers can access in asynchronous mode - jira, task tracker. The company has also now made visible the entire process of data conversion, so that the analyst can track what is happening with the data with which he works, this gives autonomy to the employee and increases his efficiency.

Andrei Malov, who lives in Israel, spoke about three weeks in quarantine on unpaid leave, a state support system. He spoke about the impossibility for Israel to work remotely from home because of the number of children in the family - there are an average of 3.5 per family.

Dmitry Garmashev said that already some time ago it became easier to hire remote employees, there is no need to come to the office and sign an employment contract, it is perfectly normal to do this in a scan by mail. For companies working remotely and having remote employees, not much that changes with the quarantine.

Nikita Uchetelev shared a life hack - in some models that are related to temporary processes (sales dynamics, for example), they added a binary feature about the pandemic period. Sasha asked a fair question whether the model would be retrained on one period, Nikita replied that in a few weeks it would be possible to conclude whether this affected the quality.

Lena Lukyanova, an employee of the Sanger Institute (Cambridge University) talked about some of the discoveries made during the period of self-isolation: the road to work and sports greatly affect her creativity. In a team where not only hard skills are important, but also software - a warm friendly atmosphere is strongly supported, the remote format does not make it possible to maintain such a connection between colleagues. Even online beer hour, at which work is not deliberately discussed, does not replace normal communication at lunch and coffee. Cambridge University understands the importance of the mental health of its employees, supports them not only now, during the isolation period, but also at any other time, providing the opportunity to work with coaches not only from this University, but also from others. Sasha Petrov confirmed that Amazon is also very worried about the well-being of employees,talked about the Mentorship program and various interest clubs for employees.

Alyona Melnikova shared the situation in Singapore, told that they have a lot of work with clients in Refinitiv, new products are being developed, work is in full swing. She confirmed that the transition to a remote site in the product team, about which Yegor spoke, did not really affect the work much, all deadlines were respected, work was being done.

Vlad Boyadzhi from Sberbank shared sore: the difficulties of interviews on skype. It was necessary to hire several people, some of them he had personally interviewed before quarantine, and some were already in quarantine. And here it is impossible to create a full-fledged impression of a person without seeing him live. From the fun - someone right in the course of the interview, using the computer at hand, begins to google the answer and is not shy. The participants of the online party immediately suggested how to deal with this))

Petya Ermakov said that he “doesn’t have enough ducks to speak out” (the duck, which Artyom Pichugin acquired in the photo below). The ManicTime application for counting time spent on different sites, in applications and documents - can be useful for analyzing your time.



In general, meeting and chatting online is also possible and we will definitely practice this in order to be able to “meet” with our nonresident and foreign graduates!