Engineering Management: An Org Structure for the Moment

Vladimir was born in Astrakhan, a steppe city on the Volga and the Caspian, and grew up in a programmer's household: in 2002 his father brought home their first computer, and his brother, ten years older, was already a programmer. From the age of ten it was Pascal, Basic and Visual Basic, games for himself and prizes at city and regional olympiads; the school's contribution, he says, was a teacher who let him skip computer science lessons for olympiad problems. He decided to leave his home town at fourteen or fifteen, and a review of universities in the magazine Igromania for 2007 or 2008 helped: admitted to Bauman and, if he is not mistaken, to ITMO, he picked MIET in Zelenograd for its green campus, built, in his words, to the design of Stockholm University. He chose the specialisation for the circuit-board view, but it disappointed him by sitting further from programming than he liked; he finished his master's with honours.

His first job paid nothing: his brother took him into the company as a general-purpose programmer — an online shop, something under Linux, a player for Android; Android is what shaped the path that followed. He found paid work in 2012, travelled two hours each way, read Habr articles on the road, and was hired, as he learned later, for a neat Android widget he had in fact only promised to write. In September 2015 he joined T-Business as, he believes, its second Android developer, and as the team grew he wanted to become a lead: management, he says, is full of things that do not yield to algorithms — communication, negotiation, the non-verbal. The team went from three people to twelve or thirteen, and his recipe for growth is short: take more initiative and more responsibility. The hardest part was the number of contexts, the rewiring of an introvert from single-threaded to multi-threaded mode; the mistake of a new lead, he says, is going back into development after six months just because it is hard, and Alexander adds that admitting the role is not for you is fine, while giving up only because adapting hurts is not.

In June 2019 Vladimir left for a project known inside the bank as project X and now as Vivid: top management wanted a bank for the European market. He arrived as the first person in mobile development, grew the team from one to twenty in a year and set up other streams, including QA; by launch engineering numbered some sixty or seventy people. He describes the product as daily banking for Europe: a debit card, account management, savings and, as he recalls, even investment products, KYC and onboarding; there were no branches, and the web at first was a landing page. Their own European licence never came, so they pivoted and launched on top of Solaris Bank's licence, the layer that talked to regulators and ran the processing. Good for speed to market, bad for economics: right after launch the team started a track to move off that layer.

Back in T-Business Vladimir had met the growth problem: headcount multiplies while productivity does not. At Vivid he picked engineers slowly and meticulously, and by launch he could hand whole areas of responsibility to strong people, which freed his time for testing and release processes across the company. The launch teams were functional: when you have a pool of features, silos work better than cross-functional teams where every business pulls the blanket its own way. After launch came product teams — crypto, investments, banking, authorisation — each with its own business, IT lead and engineers. Org structure, for him, is an instrument of the moment: functional teams win during mobilisation, product teams during growth. Later his role grew into head of engineering, and the hardest challenge was earning authority among seasoned backend and processing engineers: solving their problems helped, as did a mandate from leadership and pulling up his system design, of which mobile development has less than it might.

By the end of 2022 the company was shrinking hard, and Vladimir ran every layoff through himself, which he calls a draining exercise; together with the absence of further growth for him and his decision to stay in Russia, that settled the question, and he wrote to former colleagues, Alexander among them. In February 2023 he started as technical director of debit and deposit products: formally three tasks — draw the boundaries of the organisation, produce a technical strategy and work on time to market; in practice the brief was to do well and break nothing, because at that level nobody writes you a plan. Around fifty teams of varying maturity moved under him, and the peak load was getting to know their leads. To build authority he deliberately picked low-hanging fruit first, leaving alone root problems that would have hurt people in the moment; Alexander compares it to doctors who suppress the symptoms first. A ladder he heard somewhere and endorses: a rank-and-file developer needs intellect, a lead needs emotional intelligence, and higher up you need social intelligence and a sense of where alliances can be built.

A technical strategy, for Vladimir, is not only architecture and the untangling of legacy but a two-to-three-year view of the team, the processes, the org structure and the cost — closer to an operational strategy. Composing it is half the work: it then has to be sold to the team so they take it as their own. For the first year it lived in his head, because hiring was on fire, and hiring managers above all: there are no universal people, and every position needs its own profile. His focus now is segments — looking at a customer through their needs rather than through a product, the way a motorist buys compulsory and voluntary motor insurance and a car loan — and many teams involved report to him only indirectly, which is harder to manage than direct reports. His example of an ecosystem link is the Black card together with a SIM from the bank's own mobile operator: knowing more about the subscriber lets the protect-or-refund option catch a fraudster's call through that network. Vladimir goes to the office himself for the corridor encounters that have spawned products, but does not push his distributed team: instead the whole team, business included, gathers in one location for a week twice a year.