During January 2026, Vinh Nguyen developed two core features focused on scalable AI deployment and cost analysis. For the NVIDIA/NeMo-RL repository, he authored a detailed migration guide enabling seamless transition of training jobs to Kubernetes clusters orchestrated by Ray, leveraging NVIDIA GPUs for efficient resource utilization. In the ai-dynamo/aiperf repository, Vinh built a TCO calculator tool that integrates AIPerf benchmarking, allowing users to estimate large language model deployment costs and export results in Excel-compatible formats. His work emphasized a documentation-first approach, using Python, Kubernetes, and YAML to improve onboarding, reproducibility, and cost-aware decision-making for machine learning infrastructure teams.