Blog + media

A publishing space for writing, YouTube, references, and social distribution.

The blog is being rebuilt as a hybrid media section. Videos can carry notes and resource links beside the player, and the archive is ready for search once the volume grows.

deep-dive12:40

Low-overhead observability for Linux systems

A concept-first breakdown of why lightweight monitoring still matters for engineers working close to the machine.

This layout is designed so each video can carry its own written context, reading notes, references, and follow-up links. When your real YouTube content is added, this space becomes the companion reading panel beside the player.

LinuxObservabilitySystems

Resources and references

Video archive

YouTube videos with room for context beside the player.

Open YouTube channel
deep-dive12:40

Low-overhead observability for Linux systems

A concept-first breakdown of why lightweight monitoring still matters for engineers working close to the machine.

LinuxObservabilitySystems
talk09:15

Reading Cassandra behaviour through measurement, not intuition

A talk-oriented piece on throughput, stale reads, and what cluster metrics reveal when experiments are designed carefully.

CassandraDistributed SystemsTalk
Writing archive

Essays, research notes, and technical writing with cleaner browsing.

Profiling energy behaviour in Cassandra without losing the system view
Research note
2026-03-12·8 min

Profiling energy behaviour in Cassandra without losing the system view

What changed once energy was treated as a first-class measurement alongside throughput, latency, and consistency.

CassandraEnergyDistributed Systems
Why low-overhead Linux observability still matters
Systems essay
2026-02-18·6 min

Why low-overhead Linux observability still matters

A case for staying close to the kernel, counters, and scheduler realities when diagnosing system performance.

LinuxObservabilityPerformance
Designing experiments that explain HPC behaviour instead of just recording it
Methodology
2026-01-09·10 min

Designing experiments that explain HPC behaviour instead of just recording it

Benchmarking becomes more valuable when the setup itself helps isolate why the machine behaves the way it does.

HPCBenchmarkingMethodology
Distribution

LinkedIn and X can have a dedicated place in the publishing flow too.

Future-friendly search

Search only becomes prominent when the archive is large enough to deserve it.

The logic is already in place: once your video and article counts grow, the search and filtering UI surfaces automatically, making the archive feel more intentional and less noisy in the early stage.

Short-form and long-form video can coexist
Each media item can carry notes, links, and references
LinkedIn and X highlights can sit beside blog and video content