MUBDIUR RAHMAN

01 / Core CompetenciesEverything below is shipped, running, and mine

Browser Automation at Scale

• Playwright, Cypress, and Appium E2E automation at scale
• Structured data extraction from live production UIs
• Shared browser instances — deterministic resource usage, no leaks
• CI/CD integration: headless runs, artifacts, reliable retries

Monitoring & Observability

• Instrumented telemetry across Dynatrace, Kibana, Datadog, and the LGTM stack (Loki · Grafana · Tempo · Mimir)
• Near-realtime pipelines: back-to-back checks with no polling gaps; append-only results
• Structured logging with pino — dual-stream, plus live Express + WebSocket log viewers
• Alerting with intent — page on user impact, silence the noise

Reliability & Infrastructure

• Docker + Traefik with Let’s Encrypt; zero-downtime rebuild & deploy pipelines
• Next.js standalone production deployments on VPS infrastructure
• Server backends with rate limiting, timeouts, and SSRF guardrails
• Linux hardening: least-privilege access, containerized isolation, configuration as code

Safe & Maintainable AI

• Guardrails: prompt-injection defense, PII redaction, output validation, allow-listed tools
• Maintainable architecture: agents as state machines, versioned prompts, typed tool contracts
• Eval-driven shipping: regression gates, shadow traffic, canary deploys, automated rollback
• MCP tooling in production — AI as a force multiplier, never a demo

Backend & Tooling

• Node.js / TypeScript — zero-dependency daemons and services
• Git / CI-CD workflows, Linux administration, terminal-first operations
• Rapid prototyping with production discipline — surgical diffs, minimal deps, no magic

02 / Featured Systemmubdiur.com — the production platform behind this page

This portfolio is not a template — it’s a live production system I own end to end. A Next.js 16 standalone application behind Traefik with automatic Let’s Encrypt TLS, deployed as a single hardened Docker image with zero-downtime recreates. Eighty-plus developer utilities and an instrumented observability dashboard ship from the same repository.

The reliability layer — what actually ships

• Next.js 16 standalone — static pages, zero cold starts, sub-second TTFB
• Traefik edge with automatic Let’s Encrypt TLS — no cert rotation chores, ever
• Server-tool API guarded by rate limiting, timeouts, and SSRF checks
• 80+ client-side utilities, code-split so first paint stays fast
• Instrumented dashboard — ApexCharts telemetry shaped like production: SLOs, error rates, latency percentiles, cost per request, GPU utilization
• Live dateline and clock — no hydration mismatch, no stale content

03 / Selected Work & SystemsWhat I actually run

• Next.js 16 standalone behind Traefik with Let’s Encrypt; Docker-based rebuild & deploy pipeline
• Server-side tool backend with rate limiting, timeouts, and network guardrails
• Newspaper-style front page with a live ApexCharts observability desk — LGTM, Dynatrace, Datadog, ELK/Kibana shapes

• Operates a multi-project Docker host — the backbone for everything I build
• Hardened hosting: least-privilege access, containerized isolation, Let’s Encrypt TLS

• Long-running Node.js daemons that run unattended — no babysitting required
• Self-confirming delivery loops, structured logging, and recovery built in
• Set-and-forget systems that report back instead of silently dying