Agent Evaluation Workbench
Benchmarking multi-step tool-use reliability with reproducible test suites.
Next.jsTypeScriptPostgreSQLOpenAI API
Open project ->Research
A project portfolio of active investigations and implementation experiments.
Memory Retrieval Study
Comparing context window, vector retrieval, and hybrid memory pipelines.
PythonFAISSFastAPIRedis
Open project ->Developer Tooling Observability
Tracing task latency and failure modes across agent orchestration layers.
OpenTelemetryNode.jsGrafanaDocker
Open project ->MDX Research Knowledge Base
Designing a bilingual publishing workflow for research notes and technical essays.
Next.jsMDXnext-intlTailwindCSS
Open project ->