Image by Evgeny Bereza on LinkedIn
In mid-2021, as countries worked to reopen amid new COVID‑19 variants, governments and health systems were learning the same hard lesson: reopening borders isn’t only a policy decision, it’s a technological one. Airports needed reliable testing pipelines that could match real passenger volumes, protect public health, and keep travel moving without bottlenecks.
At Ben Gurion Airport, Israel faced this challenge without an existing COVID-19 testing infrastructure. There was no designated sampling workflow, no digital platform for bookings or results, and no diagnostic system capable of operating at a national scale. Early attempts to stand up a solution had not produced a stable, deployable system, and timelines were compressing fast as travel resumed.
The risk was immediate on two fronts. Nationally, the absence of a functioning testing pipeline threatened the integrity of health screening at the country’s main entry point. Operationally, it risked congestion, delays, and breakdowns in passenger flow at precisely the moment mobility was reopening under strict health rules. A rapid, technically sound build was essential not to fix a broken platform, but to create a working one from scratch under crisis conditions.
Evgeny Bereza, recognized for deep R&D expertise and a track record of delivering high‑complexity systems under pressure, was brought in to lead that effort. Over a 30‑day window, he directed the end‑to‑end design and launch of Checktofly, Israel’s first operational airport COVID testing platform. In parallel, he established a dedicated ML laboratory to support the second phase of the work: developing a high‑accuracy algorithm for rapid test diagnostics at scale.
The national urgency and technical significance of these projects have been highlighted in coverage by The Jerusalem Post and Globes, reflecting their central role in Israel’s pandemic‑era mobility strategy.
Starting from Zero
Bereza’s first step was not to repair an existing airport platform, but to audit the requirements and the early attempts that had failed to produce a deployable solution. At Ben Gurion, there was no established COVID testing workflow, no digital intake or results system, and no infrastructure designed to handle passenger-scale diagnostics. The risk, as Bereza framed it, was building something that looked complete on paper but could not survive real traffic.
“There was no foundation to build on,” he explains. “We had to define the architecture, the data flows, and the operating model at the same time and make sure each part could function under load.”
Rather than assembling a monolithic system, Bereza designed Checktofly as a microservices-based platform from day one, breaking the solution into independent modules for registration, lab routing, diagnostics processing, reporting, and user access. This allowed parallel development and rapid validation of each service without slowing the entire build.
The architectural choice was central to national-scale viability. By decoupling services and prioritizing testability, the platform could integrate cleanly with healthcare reporting and airport logistics, while maintaining security and performance under volatile demand. In practical terms, the framework enabled automated data pipelines, real‑time identity verification, and encrypted transmission of results capabilities required for a live testing operation at Israel’s primary international gateway.
Building a Team for Crisis Conditions
With only 30 days to deliver, Bereza implemented a focused execution model drawn directly from his prior experience leading national-scale R&D projects under pressure. Central to this was his proprietary “rapid-hire sprint” strategy—a recruitment methodology explicitly designed for crisis conditions. Instead of scaling headcount, he prioritized technical depth, selecting domain specialists who could independently own and deliver high-risk components from day one.
This recruitment model allowed Bereza to maintain a lean, high-performing unit with minimal onboarding overhead. He personally defined the system architecture, decomposed the workload into modular segments aligned with a microservices structure, and assigned each component to a dedicated lead. This ensured end-to-end ownership, reduced internal friction, and improved delivery speed.
To maintain alignment, Bereza applied his crisis-management framework: daily syncs focused on output, asynchronous documentation for traceability, and a flat reporting structure to eliminate bottlenecks. He remained the single point of architectural authority throughout, ensuring every decision aligned with the system’s functional and compliance goals.
This tightly coordinated, top-down delivery model enabled the team to operate with urgency and clarity, even under evolving health regulations. It reflected not just fast hiring, but the application of a battle-tested leadership system tailored for high-impact digital delivery in national logistics environments.
AI That Could Handle Reality
Following the successful launch of Checktofly’s operational platform, Bereza initiated a second-phase effort to address one of the most complex challenges in rapid diagnostics: automating PCR result analysis at scale using machine learning. As part of this initiative, he established a dedicated ML laboratory designed to develop and validate an algorithm suitable for national deployment.
Bereza’s methodology deviated sharply from earlier failed attempts, which had relied heavily on simulated lab data. His team instead trained the model exclusively on real-world PCR datasets, incorporating high-volume, noisy inputs to mirror the unpredictable nature of airport testing. He also introduced a micro-modular validation pipeline, an original contribution that allowed the model to be tested against isolated edge cases before full deployment.
By week three of the ML lab project, the model achieved 96% diagnostic accuracy and demonstrated robust performance across variable input data. This represented a substantial improvement over prior efforts and marked one of the first times in Israel’s health-tech sector that a live ML diagnostic tool was built and stress-tested under production constraints.
Bereza’s system was not limited to lab-theoretical success. It included integrated encryption layers, real-time ID verification, and direct interoperability with national health reporting systems. Before activation, the model and all supporting services were validated in a closed-loop environment to ensure safety and regulatory compliance.
Managing Pressure, Inside and Out
Given the involvement of multiple public and regulatory stakeholders, Bereza understood that engineering success alone wouldn’t guarantee delivery. Communication needed to be just as resilient as the code. Drawing on principles he had developed in prior high-pressure roles, he implemented a communications framework tailored for compressed timelines and multi-agency visibility.
The model was simple, intentional, and designed to eliminate noise:
- A single technical point of contact to streamline stakeholder queries
- Live dashboards displaying key metrics and system stability in real time
- Scheduled working demos to demonstrate progress and align on expectations
This structure minimized status-reporting overhead while giving external teams confidence in the platform’s readiness. Importantly, it allowed the engineering team to remain focused on delivery without creating information silos.
“We didn’t over-report,” Bereza explains. “We showed what worked, fixed what didn’t, and kept everyone aligned.”
The communication model he used, low-friction, metrics-first, and fully transparent, has since been adopted in other critical infrastructure deployments. It demonstrated that under extreme time pressure, clarity is not just a managerial asset, but an operational necessity.
Delivery, Without Delay
On day 30, the platform went live on time, without rollback or technical disruption. The Checktofly system processed its first wave of outbound passengers precisely as engineered: fast, secure, and fully compliant with the country’s diagnostic reporting standards.
Bereza personally oversaw the final validation and production handoff, applying pre-launch checklists and fallback protocols developed as part of his crisis-delivery methodology. All core components, identity verification, encrypted reporting, and diagnostics integration were deployed in a coordinated sequence.
Key outcomes at launch:
- Delivered in 30 days
- Achieved 96% diagnostic accuracy
- Integrated user registration, ID verification, and secure medical reporting
- Aligned with all national health and data privacy standards
The successful launch established a stable diagnostics infrastructure at Ben Gurion Airport for the first time. As Israel accelerated its reopening strategy, Checktofly became a critical enabler of national mobility built, tested, and delivered under one of the most compressed development timelines attempted in the public health-tech domain.
Why This Project Stood Out
Across many digital COVID-19 response efforts, the most common points of failure weren’t budget or staffing; they were weak architectures, delayed decisions, and unclear technical ownership. From the start, Bereza took a fundamentally different approach.
He applied a delivery model optimized for high-pressure environments: a microservices-based architecture, role-specific hiring, and a direct feedback loop between engineering and stakeholders. These weren’t industry-standard moves; they were part of a personal playbook developed across multiple crisis-stage projects.
“A lot of systems in this period are over-engineered or over-promised,” Bereza notes. “Ours had one job: work in the real world. That was the metric.”
The result was a diagnostic system that not only met its deadline but was structured for adaptability. As health regulations continued to shift, the platform’s modularity enabled updates without service disruptions or architectural overhauls.
This clarity of execution and stability at launch drew immediate public attention. Coverage in The Jerusalem Post and Globes recognized Checktofly’s role in enabling Israel’s safe-border strategy, particularly its combination of rapid deployment and compliance rigor.
Bereza’s core principle build less, test more is now being referenced by other health-tech and logistics teams tackling similar deployment challenges. At a time when many pandemic solutions remained on paper, Checktofly delivered a real, scalable system under live national pressure.
What Tech Leaders Can Learn
Image by Freepik
Bereza’s approach to delivering complex systems under national pressure offers clear, actionable guidance for R&D leaders managing high-stakes deployments:
- Know when to rebuild. Rather than iterate on unstable foundations, Bereza initiated a full-system audit and called for a clean-slate rebuild. This decision eliminated hidden architectural debt and enabled a platform designed for speed, testability, and regulatory alignment.
- Staff for the situation. His “rapid-hire sprint” emphasized domain-specific expertise over generalized resourcing. Each subsystem was owned by a specialist, reducing onboarding time and ensuring clarity of scope at every layer.
- Design for testability. The system’s modular design, anchored in microservices, enabled parallel development, continuous regression testing, and rapid revalidation as compliance rules shifted in real time.
- Cut communication noise. Bereza replaced layered reporting chains with direct updates, shared dashboards, and asynchronous documentation. This preserved execution momentum and reduced the risk of misalignment across external stakeholders.
Each of these methods was designed not just for performance but for resilience, critical when system failure had public health consequences. Though developed under COVID-specific pressure, the frameworks are adaptable to any project where infrastructure must scale quickly and function reliably under scrutiny.
“You don’t need to be flashy,” Bereza notes. “You just need to ship what matters and make sure it works.”
At the height of the pandemic’s logistical complexity, Checktofly served as a proof of concept for how engineering leadership, under constraints, can still meet national needs. Bereza’s work shows that, with a focused methodology, high urgency doesn’t have to mean high risk.
About the Author
Jenna Roth is a U.S.-based technology journalist covering emerging innovation in AI, cybersecurity, and public infrastructure. With over a decade of experience reporting on enterprise R&D and crisis-era digital systems, she focuses on how engineers and product leaders build what matters, especially under pressure.
