Jul 2024 - Jan 2025
Establish PoCs for multiple client project undertakings at the organisation:
- Establish POCs for Infrastructure for hosting RAG Pipelines on Kubernetes from self hosting LLM models to creating platform for ETL pipelines for the same using PyTorch, Ray, Kuberay, Qdrant and Langchain.
- Reduce GPU Inferencing cost and improve application performance. Evaluated multiple LLM Inferencing and Serving runtimes.
- PoC on implementing Keycloak for SSO and Integration with multiple API Gateways (Kong, Istio and AWS API Gateway.)

