|
|
| Domain | Tooling |
|---|---|
| AI Tooling | OpenAI API, vLLM, RAG, Embedding, Evaluation, Benchmarking |
| Backend | FastAPI, gRPC, PostgreSQL, Redis, Kafka, NATS |
| Reliability | OpenTelemetry, Prometheus, Grafana, Structured Logging |
| Workflow | CI/CD, Test Coverage, Release Automation, Contract Testing |
1) Make failures observable before making systems faster.
2) Make behavior reproducible before making demos prettier.
3) Make delivery boring: stable, testable, automatable.
- 🗣 Commented on #1820 in chen08209/FlClash
- ❌ Closed PR #1 in DestinyDevL/microfast
- AI Agent Engineering: tool-use reliability and workflow safety
- High-throughput backend services with predictable latency
- Evaluation systems for LLM applications (quality, cost, speed)
"Engineering is the art of making correct things run reliably in imperfect reality."
