Course Objective
To enable enterprises to select and deploy first-party (Gemini), partner (Anthropic Claude, etc.), and open-weight (Llama, Gemma, Mistral, Qwen) models in Vertex AI Model Garden with the right architecture across diverse use cases without single-vendor lock-in.
Throughout the course, participants explore the technical and financial trade-offs between Managed API (MaaS) consumption and dedicated GPU/TPU-backed Vertex AI Prediction Endpoint deployments.
Target Audience and Prerequisites
- Target Audience: ML engineers, MLOps/LLMOps specialists, cloud infrastructure architects, and AI platform teams.
- Prerequisites: Basic Google Cloud knowledge, experience developing APIs in Python, and familiarity with core generative AI model concepts.
Key Takeaways and Focus Areas
- Selecting models from the Model Garden catalog based on task type,
latencytargets, and data privacy requirements. - Deploying open-weight models with
vLLMor TGI-based containers onto dedicated Vertex AI Endpoints via UI and SDK. - Managing
Provisioned Throughput, dynamic workload routing, parameter-efficientfine-tuning(LoRA), and model version migrations.
Curriculum
Model Garden Ecosystem and Model Evaluation Criteria
- Capability and licensing analysis of Gemini, partner models, and open-weight model families
- Benchmarking across task complexity, context window, latency, and cost dimensions
Open-Weight Model Deployment and Serving Infrastructure
- Deploying Gemma, Llama, and Mistral models with vLLM-optimized serving containers
- GPU/TPU accelerator selection, quantization, and autoscaling endpoint configuration
Model Customization, Fine-Tuning, and Version Management
- Supervised fine-tuning, LoRA adapters, and Vertex AI Model Registry integration
- Traffic splitting, canary rollouts, and model lifecycle management
Capacity Planning, Security, and FinOps
- Capacity and quota management with Pay-as-you-go and Provisioned Throughput models
- Private Service Connect, customer-managed encryption keys (CMEK), and cost monitoring
Other Courses in This Category
-
Learn how to design and implement end-to-end tracing, telemetry, automated evaluation, safety guardrails, and IAM-based access governance in enterprise cloud agent systems.
- Live Course
- 4 modules
-
Learn how to build enterprise AI solutions, feed them with hybrid data architectures, and govern them for regulatory compliance using IBM watsonx.ai, watsonx.data, and watsonx.governance.
- Live Course
- 4 modules
-
Learn hands-on how to design and manage digital assistants, custom skill catalogs, and multi-step workflow automations integrated with enterprise applications using IBM watsonx Orchestrate.
- Live Course
- 4 modules
-
Build production-grade multi-agent systems with cyclic workflows, persistent state checkpointing, human-in-the-loop approval flows, and LangSmith observability using LangGraph.
- Live Course
- 4 modules
-
Learn how to serve open-source models like Llama, Mistral, and Qwen on vLLM, integrate them with open-source agent frameworks, and build a self-hosted enterprise AI stack.
- Live Course
- 4 modules
-
Model selection, enterprise RAG and agent development, content safety filters, automated evaluation, OpenTelemetry tracing and secure deployment on Microsoft Azure AI Foundry.
- Live Course
- 5 modules
Get in touch
