Platform-minded full-stack engineer based in Moscow. I build reliable web products, backend services, and the delivery systems around them—from transactional APIs and observability to GitOps, SLOs, and safe releases.
Core stack: Go, Python, TypeScript, React, Node.js, PostgreSQL, Redis, Docker, Kubernetes, Terraform, Prometheus, and GitHub Actions. Languages: Arabic, English, and Russian.
- Gatehouse (live dashboard, API docs) — A policy-first access operations control plane with tenant-scoped requests, role-aware approvals, idempotent writes, optimistic concurrency, a hash-linked audit trail, Prometheus metrics, a React review desk, Docker, Kubernetes, Terraform, and GHCR release automation.
- DeployLedger (live dashboard) — A self-hosted release-operations control plane with DORA metrics, idempotent deployment ingestion, signed GitHub webhooks, a hash-linked audit trail, Prometheus health signals, a React dashboard, Docker Compose, Kubernetes, Terraform, and release automation.
- SLO Forge (v1.0.0) — A Go CLI and HTTP API that compiles one SLO definition into Prometheus recording rules, multi-window burn-rate alerts, a Prometheus Operator resource, a Grafana dashboard, and a reviewable error-budget explanation. CI verifies race safety, generated drift, PromQL, the static non-root container, and high/critical vulnerabilities.
- Platform Blueprint (v1.0.0) — A reproducible GitOps reference platform with Flux reconciliation, metric-gated Flagger canaries, CEL admission controls, Prometheus SLOs, kind smoke tests, and an EKS Terraform path.
- Stockroom Ledger (live demo) — A full-stack inventory operations console with audited cycle counts, atomic location transfers, idempotent writes, optimistic concurrency, and a hardened container deployment.
- LeaseQueue (live demo) — A durable background-job service with atomic leases, crash recovery, idempotent submission, bounded retries, dead-letter redrive, and Prometheus metrics.
- SchemaSentry — A PostgreSQL migration safety analyzer with changed-migration pull-request scans, SARIF reporting, justified suppressions, and a reusable GitHub Action.
- Mini GPT Workbench (live demo) — A readable character-level transformer with vectorized causal attention, deterministic training, safe checkpoints, controlled sampling, and an interactive transformer explorer.
- Pathwise (live app) — An explainable university-program discovery and comparison product with a polished Next.js interface, search and filtering flows, and a public production deployment.
| Product | Demo | What to inspect |
|---|---|---|
| Gatehouse | Open dashboard | Tenant-scoped access requests, policy decisions, approvals, and audit evidence. |
| Gatehouse API | OpenAPI / Swagger | Live FastAPI service with health/readiness probes, OpenAPI, metrics, and auditable access workflows. |
| DeployLedger | Open dashboard | Release ingestion, DORA metrics, deployment history, and health signals. |
| Pathwise | Open app | Search, compare, and explain university-program choices in a polished product flow. |
| LeaseQueue | Open console | Queue operations, retry behavior, leases, and dead-letter recovery. |
| Stockroom Ledger | Open console | Inventory transfers, cycle counts, idempotency, and optimistic concurrency. |
| Webhook Workbench | Open workbench | Signature verification, event capture, and guarded replay. |
| Pulseboard | Open status page | Uptime checks, incidents, SLO budgets, and operational metrics. |
| Hiring signal | Project | Evidence |
|---|---|---|
| DevOps / platform engineering | Platform Blueprint | Flux GitOps, Flagger canaries, Kyverno policy, Prometheus SLOs, kind smoke tests, and Terraform for EKS. |
| Backend engineering | DeployLedger | FastAPI services, signed webhooks, idempotent ingestion, DORA metrics, audit history, and production deployment paths. |
| Full-stack product engineering | Gatehouse | React review console, tenant-scoped API, optimistic concurrency, policy evaluation, audit evidence, and a live demo. |
| Reliability / SRE | SLO Forge | SLO-as-code compilation into burn-rate alerts, dashboards, Prometheus resources, and reviewable error budgets. |
- Reliability and distributed systems: Pulseboard, Retry Lab, Incident Canvas, and Webhook Workbench.
- Policy and delivery tooling: Access Verdict, ToggleBench, and OpenAPI Impact.
- Developer tools: Tampertrail, git-hotspots, File Finder MCP, and RepoVitals.
I care about correctness under failure, secure delivery, useful observability, deterministic automation, and documentation that lets another engineer run and evaluate a project without guesswork.