Build log
Designing a Grounded Documentation Assistant
A build log on turning a documentation chatbot into a routed, source-grounded assistant that can also guide evaluation workflows.
FastAPILLMRAGStreamingTypeScript
Technical Notes
Practical notes on production systems.
These notes are intentionally short and implementation-oriented. They are a place to document how I think about backend services, release gates, and operational debugging.
Build log
RAG Beyond the Demo: Pipeline, Citations, Evaluation, and When Not to Bother
Retrieval-augmented generation as an engineering problem: what each pipeline stage can get wrong, deterministic citations, split evaluation, and how prompt caching moves the long-context break-even.
RAGLLMEmbeddingsEvaluation
Build log
Refactoring OpenTelemetry Metrics: from God Class to Per-Subsystem Modules
What a metrics-architecture refactor taught me about singletons, mixins, cardinality, and why a metric name is a public API.
OpenTelemetryPythonRefactoringObservability
Build log
An Agent Is a Loop: a Working Mental Model for Agentic Systems
Strip away the hype and an agent is an LLM in a loop that can call tools, remember things, and ask for help. Everything else is engineering around that loop.
LLMAgentsArchitectureAI Engineering
Build log
OpenTelemetry from First Principles: Three Signals, One Identity
The mental model behind OTel — signals, resources, the API/SDK split, and how trace context turns three monitoring streams into one investigation workflow.
OpenTelemetryObservabilityTracingPython
Build log
A Working Map of Computer Networking
Layering and encapsulation are the whole trick — a compact map of the five-layer model with the details that matter in practice.
NetworkingTCPHTTPDNS
Build log
What Actually Happens When You Run ssh gpu-node-02
SSH is not 'log in with a key' — it is seven distinct stages, and every error message tells you exactly which one failed.
SSHNetworkingLinuxDebugging
Build log
A Weekend of Kubernetes: FastAPI, Helm, and Breaking Things on Purpose
Deploying two FastAPI services on a local kind cluster with a production-shaped Helm chart — and why deliberately breaking the deployment taught me the most.
KubernetesHelmDockerFastAPI
Build log
Observability for AI Services: RED Metrics, p95/p99 Latency, and Tracing
A practical outline for monitoring user-facing AI services with backend-friendly signals.
ObservabilityOpenTelemetryDatadogRunbooks
Build log
Building a Production-Style Go Service on Kubernetes
A short implementation note for turning a small Go service into a deployable Kubernetes workload.
GoKubernetesHelmCI/CD