Case studies

My team was building a personal-data and consent-management platform. The project had a broad stakeholder group: dozens of engineering teams, lawyers, analysts, product leaders, and technology leaders. Decisions and newly identified risks had to reach everyone involved on a regular basis.

Requirements were not fully defined up front; my team shaped them with product partners as development progressed. Responsibility for recording decisions, risks, assumptions, issues, and dependencies was distributed across the team.

I built a knowledge base managed by an AI agent. It processed approved project sources, including meeting transcripts, team discussions, and internal documentation. At the end of each day, I reviewed the decisions and risks it had identified, together with proposed assessments of impact, probability, and the mitigation options discussed.

Over time, the knowledge base became a project copilot: one agent kept the roadmap current, while another answered stakeholder questions in team chats. The system saved around 10 hours of management time each week, and the completeness of the accumulated knowledge made it possible to hand the project to another team in one week.

The platform had been developed between 2014 and 2016 and had operated without proper support since 2017. By mid-2019 it was barely functional: payment processing failed repeatedly, while the SMS, short-code, and payment-gateway vendors could not be replaced.

The documentation had been produced as a formality and did not reflect how the platform actually worked. I reverse-engineered it: analyzed more than 20 Node.js, Java, and PHP services, reconstructed the external and internal API contracts, and reconfigured the system.

By the end of 2019, I had fully restored the platform without a maintenance budget, working as the sole specialist. The documented APIs enabled open procurement and made it possible to replace vendors on better commercial terms.

St Petersburg’s public transport authority ran a pilot competition for electronic ticketing systems on surface transport. We joined two months before the demonstration with only a minimal QR-code fare-payment prototype.

The pilot required a working system across bus, trolleybus, and tram routes: web-based ticket purchase, validation on a fixed onboard device, a mobile inspector app, and an administrator workspace with reporting, payment reconciliation, and route configuration.

I gathered requirements from the city transport authority, Passazhiravtotrans, and Gorelektrotrans, then produced the product requirements and solution architecture. I selected the validator hardware and organized its firmware development and installation. The software platform was delivered by a team under my technical leadership.

Our solution was the only entry to meet every pilot requirement, and it won the competition. The product requirements and architecture became the basis of the city’s subsequent procurement, although another company was selected to implement it. Fontanka covered how that procurement unfolded.

The city parking operator needed a subsystem for managing its parking permit registry. Operators had to request data from regional and federal authorities through Russia’s inter-agency data exchange platform (SMEV), decide on each permit, sign the decision with a qualified electronic signature, and send it to the St Petersburg public-services portal.

I delivered the subsystem end to end as the sole developer in six months. Working with the customer’s specialist unit, I gathered and completed the requirements, integrated eight data services across SMEV2 and SMEV3, implemented decision signing and delivery, and prepared user and technical documentation to GOST 34 standards.

The subsystem passed testing and pilot operation, was accepted by the customer, and remained part of the city parking platform for several years. Its database and software component are registered with Rospatent.

In early 2020, amendments to Russia’s Code of Administrative Offences were being prepared to let paid-parking operators obtain vehicle-owner data from the Ministry of Internal Affairs. The entire process had to be launched: requesting the data, issuing and delivering each notice, and confirming payment through the treasury system.

There was no development budget. No integration with the Ministry existed, and the existing connections to Russian Post and the State Information System for State and Municipal Payments (GIS GMP) were outdated.

I analyzed the process, agreed a data-exchange format with the Ministry, and implemented integrations with Russian Post and GIS GMP using the new protocols. Because the Ministry’s online service was unavailable, we organized the exchange through registries on electronic media. I also built a service for payment control, record keeping, and reporting.

The first 20 fines were issued the day after the amendments took effect. By the end of February, the platform could process up to 1,000 notices per day, deliver them in paper and electronic form, and verify payment through GIS GMP.

The center operated 24 automated city and park-and-ride parking facilities supplied by three vendors. Because the maintenance budget was insufficient, only four facilities could operate fully without attendants at the start of the year.

I studied each vendor’s hardware and software, created preventive-maintenance procedures, and trained two internal teams to handle routine and emergency work. The existing contract budget was redirected to consumables and spare parts. Equipment from one facility that could not be restored under the constraints became a source of parts for the other sites.

Within three months, 20 of the 24 facilities were fully operational without any budget increase.