SLOs over vibes
Translate user impact into service objectives and error budgets that guide engineering trade-offs.
Site Reliability Engineer · Production Systems
Keep the signal.
Lose the noise.
An SRE focused on resilient platforms, useful observability, safe delivery and calm incident response. I turn operational risk into systems teams can reason about.
SERVICE
HEALTH
HEALTH
Telemetry
SLOs
Deployments
Incident response
01 / Reliability ethos
Calm systems
scale better.
Reliability is not a heroic act at 2 a.m. It is the compound effect of good defaults, measurable objectives, fast feedback and systems designed for failure.
Signal over dashboards
Design telemetry around questions responders actually need answered during normal operation and incidents.
Remove repeat toil
Turn recurring operational work into safe, documented automation with clear ownership and rollback paths.
Blameless by default
Use incident reviews to improve systems, interfaces and detection rather than hunting for individual fault.
02 / Operations
Production
under control.
Selected operational stories show how I reduce blast radius, shorten recovery and make the next incident easier to handle.
P1 · 14 MIN
Checkout latency spike
Traced a cross-region dependency regression, introduced saturation alerts and redesigned retry behaviour.
P2 · 28 MIN
Queue backlog storm
Added consumer lag SLOs, adaptive capacity and a replay-safe recovery runbook.
P2 · 19 MIN
Deploy health regression
Built progressive delivery checks around golden signals and automated rollback conditions.
03 / Expertise
The reliability
toolbox.
Expertise is represented as signal clusters rather than artificial percentage bars.
Cloud and Kubernetes
Design resilient workloads, capacity boundaries, autoscaling and failure-aware platform patterns.
Metrics, Logs and Traces
Build telemetry that connects service behaviour to user impact and actionable operational signals.
Safe Deployments
Progressive delivery, release health, rollback strategy and automation for confident change.
Chaos and Failure Modes
Model dependencies, test recovery paths and make failure behaviour explicit before production finds it.
Response and Recovery
Lead calm incident coordination, technical diagnosis, communication and durable follow-through.
Go, Python and IaC
Automate operational workflows with maintainable code, infrastructure as code and strong defaults.
04 / Selected systems
Reliability
in practice.
Case studies focus on the operating model, failure surface and measurable outcome, not a list of tools.
Multi-region platform with graceful degradation
Reworked service dependencies, traffic policies and recovery automation for a customer-facing platform serving peak traffic across three regions.
99.95
target availability
37
% less incident impact
One signal model for 40+ services
Standardised golden signals, ownership metadata and alert quality so on-call engineers could distinguish symptoms from causes.
Release guardrails that know when to stop
Progressive rollout checks tied to latency, errors and saturation with automated rollback for unhealthy releases.