Data Engineering: From a Rocket Plant to Amazon

Dmitry Anoshin introduces himself like this: born and educated in Moscow, a design engineer by training, three years at the Khrunichev plant that builds Proton rockets. By 2010 it was clear that mechanical engineering would not feed him: on paper he headed a machining department, in practice he earned around twenty thousand roubles a month, while a university friend who had landed a project manager's assistant job in IT consulting earned fifty. Then begins the plot repeated throughout: he finds opportunities himself. He and a friend worked through the plant's internal phone book, calling every department with any hint of computers, and settled in the automation department, whose head reminisced about the eighties and the automated trolleys in the workshop while his staff watched TV series and played games online. Dmitry then wrote to every company on the RBC ranking of IT firms: two replied, one refused, the other offered installing internet cabling with no promise of a later move onto the IT staff.

What worked was a forgotten intern application form: a bank invited him to an interview, only his second ever and so frightening he could barely speak. All he knew about databases was Microsoft Access, and he agreed to work unpaid for three months. His parents did not understand why he was leaving the plant; his wife backed him. At the bank he landed on Business Objects, whose training courses came in English — hence his insistence on reading documentation in the original, or else one is left programming in 1C. Alexander offers the counter-example: when Russian was added to an internal tool's instructions, its NPS among developers rose noticeably. Dmitry had wanted to leave the country from the start: he spent summers volunteering in Italy, France and Germany, and once at the bank he searched an Italian job site for Business Objects. Work visas came to nothing in Europe, Asia or Dubai; residency was offered by Canada and Australia — Canada it became.

He found the job while still in Moscow: a CV with a Canadian address and Canadian IP telephony forwarded to his phone, the plan being to slip past the recruiter and reach the hiring manager. While the papers were processed he lived nearly a year in Montenegro, and his first job was an insurance company in Winnipeg: a city of, by his reckoning, about a million, snow in May, minus forty in winter, colleagues of ten years' tenure whom his motivation only got in the way of, and the day he, a senior specialist, was sent home to change his jeans for proper trousers. The Amazon interview fell in April: snow in Winnipeg, flowers and the ocean in British Columbia. Formally he was hired as a BI developer, but inside Amazon's system the opening was for a data engineer — in 2016 the role was only taking shape. His approach is to step outside the job description: if you build a dashboard, work out how the warehouse behind it is built, what loads data into it and what the alternatives are.

Dmitry places himself firmly on the individual contributor track: in the West an IC earns almost as much as a manager, with fewer meetings and better job security. Alexander describes the same role from a manager's chair — a staff engineer designs the concept of a solution and acts as a technical coordinator, and the position is hard to explain because it is neither boss nor executor; it earns its keep at the seams between domains, which sag because every team mentally leaves the hard part to its neighbours. The most valuable thing about five years at Amazon, Dmitry says, was the culture: leadership principles there are used not under duress but because people believe they made the company successful. His own are ownership, deliver results and bias for action — better to act and apologise than to wait for permission. At Microsoft, he says, there was nothing of the kind, and the single thing he took away was the growth mindset.

Growing to senior is easy, Dmitry says: you work, you learn, you change jobs. After that comes everything unrelated to technical skill — soft skills, politics, bureaucracy. In five years at Amazon he was never promoted, with an annual indexation of 2%, and he names his own mistake: in a small team serving analytics for a hundred and fifty people he helped everyone indiscriminately instead of working with the three directors his promotion depended on. Hence the advice to write down goals and the route from A to B with your manager on day one. Alexander adds that a promotion package can also be assembled after the fact, from two years of artefacts, but at staff level they will still ask for meaningful scope, impact and leadership. By his fifth year Dmitry realised he was going through the motions with no result: the burnout came from expectations he had invented himself, and a promotion plus 50% at Microsoft changed nothing.

He started his Telegram channel in 2018 around his own Snowflake and Tableau lab at a Moscow marketing analytics conference, where he spoke on Amazon's behalf without asking Amazon; by the end the channel had 600 members. Then came a request for a course, a refusal to hand over the syllabus for a hundred thousand roubles and the conclusion that paid schools had turned into businesses with good marketing — and, in response, the free Data Learn project with three years of YouTube videos. His new English-language community has a paywall, not for the money but so that people join deliberately; Alexander confirms it from his own experience: he stopped abandoning courses halfway once he started paying. Books follow similar arithmetic: an advance of around three thousand dollars split between co-authors, royalties in pennies, and six to eight months of writing during which the product moves on, so more co-authors simply means shipping sooner.