Get best of the week

How an engineer grow into techlida

Who are the team leader, architect or QA and what they do, in IT, just about everything is imagined. But with the understanding of who the technical expert is, what he is responsible for and how to become one, difficulties arise. We conducted dozens of interviews with specialists from large companies and learned that this is an engineer who initiates processes: connects people and tools with the goals of the organization. He takes the initiative and responsibility for the technological development of the product and cares for the quality of technical solutions. Moreover, quality is not only testing, but architecture, design, engineering practices and experiments, work with the technical debt and technical improvement of the company as a whole.



We also found out that there are many conferences for techlides. But almost all of them focus on tools, not engineering practices and processes. That is why we launched the new TechLead Conf 2020 Online conference - for those who would like to become a technical expert and understand what quality is. 

At TechLead Conf 2020 Online, the secondary question is “What technical tool was used to solve the problem?”. This conference is for those who strive for the quality of technical solutions and take responsibility for the technological development of the product. From June 8 to 10, we will study the experience of implementing and using practices, technology and process management in the company. We will tell you more about the program and what we will talk about at the event.

Short program


The TechLead Conf 2020 Online program goes from discussing the development of techlide to putting DDD into practice and consists of several blocks.

  • Map technical development . There is still little understanding of who this is and what it does. And the questions of how to grow up to techlide and what it should be able to do are asked even less often, so in the first block we will discuss who it is and how to become one. 
  • . — , -, « ». , — : , . « ». , MVP.
  • Delayed effects of engineering practices . In working with code, feedback is fast: wrote, tested, deployed, works. But in the technical world, the result of his work is noticeable only after months. Therefore, we added reports on all stages of the life cycle of engineering practices: the emergence of an idea, MVP, the prevention of errors and the measurement of results after a successful launch on real cases.

Also discuss: 


We will tell in more detail about each block and reports in them.

Techlide Development Map


Tehlid is the role of an engineer who manages processes. Typically, these are engineers at least senior: developers, architects, automation, SRE, less commonly CTO. Sometimes they can be a team leader. But the team leader builds teams, manages people and their development.
Tehlid builds processes: makes technical decisions that affect product development in the face of uncertainty.
Therefore, the conference will not have reports on people management and motivation, but only topics on technology management, technical leadership and the construction of engineering processes. And the very first thing to learn is how to become a good techlide.

The success of the company depends on the availability of strong specialists. We have already described how techlide differs from other professions, and Vladimir Gorovoy , Data Science Product Manager in Yandex.Verticals , tells Vladimir Gorovoy about the difference between a good techlide from other professions . From the report “ How to Become a Good Techlide»Find out where to start our development, what skills and qualities to pump. Vladimir will share his rich experience in participating in the creation of Yandex.Travel, Yandex.Real Estate and Yandex.Market projects to illustrate the thesis.

The stronger the skills, the easier it is to cope with your tasks. But the pain does not go anywhere - they are about the same for all techlides.

  • Write code or engage in a technology development strategy for the product and team?
  • Solve complex technical problems yourself or delegate?
  • How not to get torn between writing quality code and rolling out features?

These and other conflicts will be sorted out by Evgeny Korytov . In the report " Problems of technical technology and how to solve them? "Eugene will tell you how to cope with the tasks and problems of leaders, with the help of a framework that" solves problems. "

Combining business and development


Maintaining high quality code and making the right technical decisions is not the whole job. You also have to constantly prove to business and customers the need to invest time and energy in architectural and technical tasks. Alexey Deryushkin from Better Life Company knows from experience what it is like: 15 years of team leadership and 5 years of consulting experience. In the report “ How to“ Sell “Technical Tasks to Business ”, he will show how to conduct a dialogue with business in order to make cool features and not to forget about quality using life examples.

The struggle between business and development is a standard problem in IT projects. Often the business does not have TK, but only “ideas” and requests. This leads to incorrect decisions that have to be corrected for months or maintained for years on crutches. How to find a balance between development and business, Arthur Dementyev will share in the report “ Between Two Lights: Development and Business ”. Arthur in IT since 2009, on the example of stories from practice, will illustrate different approaches to the implementation of MVP features.

When the business and development have agreed, it is time to proceed to the next stage. Now techlide has several months to introduce a new product. Most often, in such situations a minimally working hypothesis testing product (MVP) is created. Immediately after testing, the prototype code is thrown out and the application is written "as it should." But what to do when the test launch was successful, and the real users already live in the "craft"? We learn from Maxim Arshinov from the report “ How to launch MVP and not turn it into a technical debt ”.

We select and implement engineering practices


It’s always difficult to implement something new, the same MVP, PL or framework. The novelty may turn out to be “raw” and not live up to expectations. How to make the right choice and “ Introduce a new technology and not waste all polymers, ” says Pavel Mineev , team lead from Rocketbank.

For the introduction of new products, it is optimal to use the governance as a code approach. With this approach, all rules have their own life cycle, they are subject to testing and are no different from a regular software product. From the report of Alexander Tokarev “ Governance as a code: how to comply with development standards and not slow down the delivery of features»We learn how to apply this approach: how and what to check during the software development process, how the approach allows you to develop more secure and high-quality applications. 

When the standards are implemented, it's time to test them, for example, on something large-scale - to create a technology platform. MTS is a large IT company implementing projects from telemedicine to IoT. Each new project stimulates demand for the following and reduces the cost of their creation. This is only possible through the implementation of best engineering practices. But there are difficulties: hundreds of teams with different levels and processes, legacy, the need to "sell" ideas to business. How do companies cope with these tasks, we learn from the report “ What should we create a technology platform?” A step-by-step guide from idea to implementation . ” Tells Secrets Philipp Bocharov- Project Manager for Development at MTS IT.

After the selection and implementation of engineering practices, work is just beginning - you need to evaluate the result. This will help metrics: it is important to understand not only what is happening with the infrastructure and hardware, but also how each feature works, to find bottlenecks and remove them in time. The report “ Set up monitoring and what next?” » Mikhail Mazein will share metrics on the example of ManyChat - a platform where a million active businesses communicate with 800 million of their customers. What to consider:

  • how to work with metrics under heavy load and with regular releases;
  • which ones to monitor first;
  • how to build a rapid response process and learn about problems in the service before users.

Platform teams


Back to the platforms. Several different teams work on their development and support. They are responsible for their zones, but there is no one responsible for everything — there are “through” pains. These problems are solved by platform teams: they create the infrastructure for developing applications and their work on production, help to work faster and better, are responsible for everything. How to create such a team and apply product thinking, says Dmitry Vishin , head of the goods.ru platform development group, in his report “ Platform teams are important. Why? "

Creating a platform team is not enough. You need to be able to not break it down before someone starts using the platform. On this way, evil raccoons can be found. Yes, it’s raccoons. Where do the raccoons come from and how are they related to the stabilization of the platform team ? We learn from the lead developer at MTS, Elizabeth Golenok, from the report “ Platform Team and 4 Evil Raccoons ”.

The reports will be supplemented by a round table discussion “ Platform teams: benefits or harms ”. During the round table, Philip Uvarov (Spotify) and Andrei Alexandrov (Mafin) will discuss several issues.

  • Why are these commands needed and are they needed at all?
  • Why has it become fashionable to create them?
  • Is there any use for them or is it hype?
  • « », ?


Despite all the engineering practices and the help of the teams, the techlide writes the code. How to write in such a way that the code is readable and supported, and not rewrite everything in a year? Two reports will answer this question.

The first is “ How to Write Readable Code, ” by Grigory Petrov , Head of Developer Relations at Evrone. Gregory organizes development, conferences ( Moscow Python Conf ++ ), hackathons, generalist and amateur neurophysiologist. As a result, the report will have a lot of neurophysiology, cognitive and social intuition. But the main thing is that Gregory will tell you where the code complexity comes from, why it cannot be removed and how to live with it.

The second is the report “ Balance of contradictions. Selection of the best practices in the code and in the team » Gleb Lobastov. Gleb is a technical advisor and leader of the development team at OneTwoTrip with 10 years of experience. The report will share approaches to writing "good" code - understandable and convenient to support, and will answer several questions:

  • what to consider when implementing best practices from the point of view of the project and the team;
  • the main enemy of good code and how to deal with it;
  • contradictions in the practice of writing good code.

All this with examples, with a set of principles and practices for writing code that you can be proud of.

Legacy and refactoring


The code topic, or rather the old code, will continue in the block on legacy and refactoring. Many are familiar with static analysis as a convenient tool. But sometimes difficulties arise, for example, when the project has a huge database of legacy code. When statistical analysis finds errors, what to do with them? How to balance the corrections of old errors and catching new ones? We learn from the report “ How to fix hundreds of errors in legacy code and not die (for example, Unreal Engine 4) ” by George Gribkov .
You can refactor not only the code, but also the architecture, infrastructure and processes.
Any long-lived IT company is faced with a slowdown in production processes. It is provoked by many factors, for example, the complexity of technology and an increase in the number of employees. This leads to the fact that coordination is delayed, no one bears responsibility, and systems become fragile. Lev Goncharov (T-Systems) in his report “ Agreements as Code: how to refactor processes and not break down ” will share stories from 14 years of experience that will help accelerate infrastructure processes and make them explicit.

After refactoring the infrastructure processes, you can proceed to standardize it. For example, get rid of the "zoo" of technology. How to do this on the example of the experience of standardizing the infrastructure of one particular large application, Ilya Mitrukov will tell- Infrastructure Manager (Technical Information Security Officer) of the Deutsche Bank Technology Center. 

The report “ Not the gods burn pots. Standardization of infrastructure ”there will be nothing about the upgrade of technologies, innovative solutions, technological“ Cosmos ”or CI / CD pipelines. There will be only infrastructural life of a couple of years of projects, minimizing costs and supporting business development.

From code, processes, and infrastructure, let's move on to technology refactoring. How do I translate a project that commits 70 people a day to React and TypeScript so that no one notices? We’ll ask Yandex, or rather Evgeny Dashkevich , the head of the Yandex group. In the report “ How to move to a new technology so that 70 developers do not notice anything"Eugene will share the translation history and the reasons for updating the technical stack in the project, which renders millions of different combinations of search results per day.

DDD, Event Storming and Knowledge Management


In this section of the conference, we will talk about design, using the approaches and practices of DDD - Domain Driven Design (domain-specific design). Often it is abandoned due to the fact that this is a methodology without clear indications of what and how to do. But Raiffeisenbank has been using DDD practices in various projects for 5 years to decompose the system into microservices, communicate with customers and within teams, and create applications that do not resist new requirements. How to apply the approach, which practices to use and avoid mistakes, we learn from the report of Konstantin Gustov " How to tame DDD ".

There are many practices in DDD. One of them is Event Storming. It facilitates further work in the field of DDD and microservice design. When creating a system on microservices, you can easily create a distributed monolith. Event Storming does not protect against this by 100%, but it can significantly reduce the risk. About how, with practical examples, in the report of Sergey Baranov (ScrumTrek) " Modeling microservices using Event Storming ".

When we figured out what the techlide is doing, how to develop and implement engineering practices, we move on to the storage, management, knowledge sharing and tracking of technical solutions in a team. For example, knowledge management is needed when developers make similar functionality in different parts of one large project. They spend many times more time and resources on this than if they combined their efforts.

Ilya Kashlakov heads the front-end development department of 50 people in Yandex.Money. With so many developers, it’s vital to share knowledge and keep an eye on architecture. The report “ Logic Review - as a tool for making complex technical decisions"Ilya will talk about this tool: how they came up with the Logic Review, what metrics they collected and how they determined the success of this process. All this with examples of problems, a description of the changes that have occurred in the process from the start to the present day.

For the implementation of projects we need a lot of documentation. To store it use, for example, lightweight markup languages: Markdown, reStructuredText, Asciidoc. They are easy to write, and files are conveniently stored in the repository. At the workshop “ How to publish Markdown and RST? Review of modern documentation tools ”we’ll talk about how to apply them to techlides:

  • Konstantin Valeev (Rostelecom IT) will share a way to create customized PDF and HTML from Markdown sources;
  • Semen Faktorovich (documentat.io) will talk about the “Swiss knife” of Pandoc and how to defeat the DOCX generation with it;
  • Nikolai Volynkin (Plesk) - how to generate huge HTML portals using Sphinx-doc.

Three speakers will share their experience and everyone will be able to ask them their own question on the topic.

TechLead Conf 2020 Online for those who want to grow into techlida


Conference TechLead Conf 2020 Online for tehlidov and those who want to become engineers, developers, team lead, of QA, development managers. Even if you are not a techlide yet, come to the conference and collect instructions for you on how to become one - a map of the technical competencies. 

The whole conference will be held in a new format - online. Thanks to this, we added more content in three days of the event than offline: more than 30 reports, lightning talks (short reports with answers to questions), OST for the exchange of experience, a round table and various networking formats. We have prepared a schedule for the program - look, mark your favorite reports or master classes in the calendar so as not to miss.

A new format - new prices, so that even in these strange times you can continue to develop and maintain contacts with colleagues from other companies. Book tickets - the price of 5900 for individuals will help to keep abreast of new products in the industry or get a new profession.

All participants of the online conference in the Personal Account can offer their question for discussion, request help in a work task or start an interesting and relevant discussion. There you can immediately vote on exciting topics. Authors of the best ideas will receive a free ticket to the selected conference, where the proposed discussion will be organized.

29 18:00 -. , maturity model.

. . live-, Spring Cloud Contract Pact. , .

More articles:


All Articles