Building Smarter Software, Greener Futures
Robert Markovski — Founder & CEO, Prilog
Robert Markovski is the Founder & CEO of Prilog, an AI-powered software maintenance agent built to close what he calls the "remediation gap" — the costly distance between detecting a production issue and actually shipping the fix. In a conversation with Myles, Markovski unpacks why observability alone is no longer enough, how Prilog fits into existing engineering workflows, and why a partnership with reforestation platform Evertreen is central to the kind of company he is building.
Engineering teams today have no shortage of visibility. Logs are ingested, traces are correlated, anomalies are surfaced on dashboards in real time — yet the moment something breaks in production, a senior engineer is still paged at 2 a.m. to fix it by hand. Markovski sees this gap between knowing and fixing as the defining inefficiency of modern software operations.
Prilog sits on top of existing observability stacks — including Datadog, SigNoz, Grafana, CloudWatch, ELK, Prometheus, and major cloud providers such as AWS, GCP, and Azure — and does what no monitoring tool does: it actually fixes things. The platform ingests logs, alerts, and incidents, clusters recurring issues, maps them to the responsible code, and generates reviewable pull requests. The workflow moves from alert to root cause to shipped fix, automatically, without replacing the human review step.
Critically, Markovski is clear about what Prilog is not. It is not another dashboard, not another alert layer, and not a generic AI code-generation tool. It is, in his framing, the missing remediation layer.
Prilog operates through a three-stage process. First, it detects — grouping noisy logs and traces into the subset of incidents that are actually worth addressing. Second, it maps — pinning the exact repository, service, deployment, and function behind each incident. Third, it fixes — drafting a verified patch and opening it as a pull request, ready for a human engineer to approve and ship.
Underpinning this is what Prilog calls a Production Remediation Graph: a continuously updated model that connects incidents, repositories, service dependencies, review feedback, and recurrence history. Every approved fix makes the next incident faster to resolve, meaning the system learns from the way each specific engineering team works — how their systems fail, how their team reviews code, and which remediations actually survive deployment.
The distinction Markovski draws is meaningful. General-purpose AI models can generate a patch from a prompt, but they have no awareness of a team's production history, dependency graph, or review preferences. Prilog, by contrast, is purpose-built around production context. It does not train on customer data for shared foundation models, and it is designed to remember the specific failure patterns of each customer's systems across services and deploys.
Markovski founded Prilog after running mozok, where he experienced first-hand how much engineering capacity quietly disappears into repetitive firefighting. That operational background shaped his conviction: engineering teams do not have a visibility problem, they have a remediation problem.
Alongside its core product, Prilog has formed a partnership with Evertreen, a digital reforestation platform that enables businesses to sponsor tree-planting and reforestation projects worldwide. For Markovski, this collaboration is not a marketing exercise — it reflects a deliberate choice about the kind of company Prilog is becoming.
Evertreen works directly with local farmers and forestry experts, and is a VERRA-certified partner trusted by more than 1,000 companies globally. Every tree planted through the platform is tracked with GPS coordinates and satellite monitoring, giving partner companies real-time visibility into the number of trees planted, estimated carbon offset, and the specific restoration projects their contributions support. That emphasis on transparency and accountability mirrors the values Prilog applies to its own product: nothing is a black box, every action is reviewable.
When an organisation chooses Prilog, the partnership with Evertreen means that choice carries a tangible environmental dimension. The mechanism is transparent and measurable — consistent with Prilog's broader belief that long-term value is built through actions that can be verified, not just claimed.
Markovski's stated ambition is to create a world without software bugs. That is not a product roadmap milestone — it is the organising principle behind every decision, from the architecture of the remediation engine to the choice of sustainability partners. The vision is of software that heals itself: production systems that learn from their own failure history and close issues before they consume another engineer's weekend.
For organisations evaluating technology partners, Prilog's proposition combines near-term operational impact — reduced mean time to remediation, less repetitive toil for senior engineers, lower technical debt — with a longer-term signal about how the company is run. The Evertreen partnership is one expression of that signal: a commitment to building something that lasts, and that takes seriously what it leaves behind.
Comments
No comments yet — be the first to share your thoughts.