S1LIVE
YEAR2026
STACK10 PKG
STATUSACTIVE
LOCATIONREMOTE
PACKAGES10 installed
- ▸Constructed core inference stack for sovereign multilingual LLM deployment, packaging custom Docker images via BuildKit and serving models as dedicated pods alongside vLLM/TGI on enterprise hardware
- ▸Deployed FastAPI serving layers for open-source multilingual models, orchestrating containerized inference servers optimized for enterprise workload throughput
- ▸Instrumented observability with Langfuse, LangSmith, and Grafana monitoring e2e latency, token usage, and TTFT; RouteLLM router selected optimal model 80% of the time
DockerBuildKitvLLMTGIFastAPILangfuseLangSmithGrafanaRouteLLMPython
View OPEA contributions (arpannookala-12)