From Lead Back to Senior: Rebuilding Influence from Zero

Andrey starts at the beginning: he comes from Cherkessk, a small town in the North Caucasus, and went to Stavropol to study business informatics — pointless studies, by his account, while the best decision was finding a company where he started programming and earning money instead of attending lectures. Then Moscow: the local internet provider wanted to go nationwide and moved its development team to the capital to run every city from one point, but it never came together. Andrey chose not to return to Stavropol, went job hunting, wanted Rambler and did not get in, and in 2016 landed at Tinkoff as a middle developer. The team there was finishing a new version of the site on react and ssr: the deadline slipped by half a year, bugs piled up, and beta users asked for the old version back.

Alexander recalls joining the same company in 2016 and apologises: he had attributed some of that platform's decisions to Andrey although Andrey had inherited them, so the complaints were unfair. Andrey explains that the stack was new, nobody could write react — the team had been an Angular one — and the first version was put together any which way. He joined a product team doing payment functionality, but infrastructure pulled him: the build system, dependency updates, how the application was arranged. The platform team's lead invited him over: fifteen-plus people on the project and plenty of work. Andrey became a senior, then the lead of that team, and inherited a monolith of many hundreds of thousands of lines and forty developers — hard releases, any change able to break everything, teams demanding a rewrite or another stack. His job was to make it a place where developers stop complaining; complaints did go down.

Alexander then recruited him into the platform teams of the advertising business, with similar infrastructure problems plus a goal of attracting customers; that was already a manager-of-managers role, around twenty people, though Andrey does not remember exactly. Alexander adds context: the unauthenticated part of the site and the personal accounts kept growing, a server-driven UI engine and its forms lived next door, and what was expected of him was not tools for their own sake but tools product teams would adopt. A year later he moved to the company-wide platform team and ended his Tinkoff path as the frontend lead: the Tramvai framework and the approaches to modular applications travelled with him and became the de facto standard. Back as a middle developer he had already concluded that technical skills arrive on their own, while soft skills must be trained hard — some leads were manipulative, and refusing to work with them was not an option.

Boredom pushed him into projects beside the frontend. The first was Statist: client application metrics had to be collected, so Andrey built the zeroth and first versions and laid the architecture; Alexander confirms it is alive, product analytics now sits in the data platform tooling, and it is being married to the experiments platform, since the methodology of hypotheses and metrics is needed company-wide. The second was a service that spins up Cassandra clusters with almost no manual steps; his infrastructure chapter started there — Kubernetes, hardware, data centres and a pile of unfamiliar terms — with about seventeen people on his team. The third was a wrapper over the cluster that folded deployment complexity into a few commands: meant to be temporary, built to die once a proper solution arrived — but the proper one was a long time coming and the wrapper took half of all deployments. Andrey also launched the growth and promotion process: turning "this is a bad idea" into "fine, not bad" took a series of meetings and hundreds of people to convince; a new generation uses it now, often without knowing who devised it.

In February 2022 — he does not remember the date — Andrey left Russia for Cyprus, worked at Tinkoff a while longer and, living in London by then, moved to Booking: a platform team again, but as a senior, a deliberate downgrade from a managerial role. There he developed internal Node.js products and worked a lot with AWS, concluding that he dislikes the public cloud: too low-level, no high abstractions, easy to shoot your own legs off and then pay for it every month. He calls the atmosphere pensioner-like: the bar next to the office was packed at three on a Wednesday afternoon, while burnout surveys had everyone reporting maximum exhaustion although, by his observation, people were mostly chilling. The free time went into databases and networks — that is where he understood how Cassandra works inside — and he started a blog to improve his written English. On LLMs he says the model generalises everything and strips out details, and Alexander calls the result a worn-down sliver of thought: write in your own words and leave the model the grammar.

Andrey is now a product developer writing react at an American company he does not name and will not discuss; his infrastructure experience sits unused, and his grade is an ordinary senior level five. Two things block further growth: Russian companies on a CV read as a lack of local experience, and the English he began learning some six years ago makes confident pitching across three hour-long behavioural interviews hard. Hardest is the lost influence: at Tinkoff he was known and listened to, whereas now his opinion is one of six with a level seven and eight nearby, and the right to be heard must be earned again. He is blunt about school: a C student with dyslexia and twenty mistakes per dictation — but that grew the abstract thinking that lets him hold a whole large system in his head. The closing advice is unglamorous: he got into IT because a friend needed a website, and direction, steady work and luck did the rest. There is no magic pill, Alexander sums up, and the participants note that FAANG-type companies hire not the smartest but those ready to work hard and pass every challenge.