Tools2026
01
readme-gen
A zero-dependency Rust CLI that scans a repository, renders a bilingual README, and fails its own check when a documented number drifts.
Rust 2021CLISHA-256+2
Case study
Portfolio / 2026
Systems that stay understandable as they grow.
I'm Miles Li, a platform engineer working across Kubernetes, cloud infrastructure, distributed middleware, observability, and automation. I turn fragmented infrastructure into systems that behave predictably.
The system is the core; I am its architect.
- 38
- public repositories
- 6
- case studies
- 84
- photographs
Field notes
84 frames
01Selected work
Systems, benchmarks, and small sharp tools.
Six projects that best represent how I work: make the boundaries explicit, measure what matters, and keep the result operable.
Platform2024–2025
02
Containerized Environment
Reusable Docker dev containers for six toolchains on macOS and Windows, plus Kubernetes manifests and a Kafka/KEDA autoscaling walkthrough.
DockerComposeKubernetes+3
Case study
Performance2025
03
Computing Benchmark
A C++23 benchmark for blocked matrix multiplication that reports GFLOPS, speedup, and parallel efficiency across machines.
C++23OpenMPCache blocking+2
Case study
Performance2025
04
Thread Pool Study
A team study of nine matrix-multiplication variants — FIFO, LIFO, async, TBB, OpenMP, and BLAS — measured with Linux perf on three CPUs.
C++OpenMPTBB+3
Case study
Observability2025
05
Grafana on Kubernetes
Kubernetes manifests and a bootstrap script for Grafana with a 15 GiB persistent volume, probes, resource limits, and an optional ingress.
KubernetesGrafanaPVC+2
Case study
Systems2026
06
TCP Proxy Prototype
An early C++20/Boost.Asio TCP proxy prototype: two coroutines per connection, an 8 KiB buffer per direction, hardcoded routing.
C++20Boost.AsioCoroutines+1
Case study
02Capabilities
What I work with.
The tools change; the approach does not. Define the boundary, automate the repeatable part, then make the state observable.
Cloud & Orchestration
Kubernetes-first platforms with declarative delivery and reproducible environments.
KubernetesEKS / ACKDockerHelmTerraformGitOps
Distributed Systems
Stateful middleware operated in production, including cross-region migration work.
KafkaZooKeeperRabbitMQReplicationZero-downtime cutover
Observability
Turning runtime state into evidence: metrics, logs, dashboards, and alerts.
PrometheusGrafanaELKSplunknode_exporter
Automation & Tooling
Reducing operational entropy with pipelines, configuration, and small sharp tools.
GitHub ActionsJenkinsGitLab CIAnsibleRust / C++ / Python
Working principle
“The system is the core; I am its architect.”
01Eliminate what is unnecessary.
02Isolate what must coexist.
03Automate what can be declared.
04Observe what cannot be assumed.
03About
I build systems that stay understandable as they grow.
Hengyi LiPlatform Engineer · Cloud-Native
I work at the intersection of Kubernetes, cloud infrastructure, distributed systems, observability, and automation. Most of my time goes into turning fragmented infrastructure into reproducible, operable systems.
That usually means removing accidental complexity, drawing explicit boundaries, and making failure modes visible before they become incidents. The same instinct shows up in smaller projects: a README generator that refuses to lie about its numbers, a benchmark that reports its own scaling efficiency, a proxy built to understand the layer below.
Away from the terminal I photograph landscapes and quiet in-between moments. The gallery is the other half of the same practice: paying attention to what is actually there.
04Field notes
Photographs, mostly landscapes.
The other half of paying attention: light, weather, and places that do not need to be optimized.
