Tag: AWS
-
Amazon SageMaker JumpStart, Foundation Models and Private Hubs (AWS Gen AI Series, Part 18)
SageMaker JumpStart gives you open-weight and proprietary foundation models on your own SageMaker endpoint. Here is how it differs from Bedrock, what it costs, and when the instance bill is worth it.
-
Amazon Bedrock Model Distillation, End to End (AWS Gen AI Series, Part 17)
Amazon Bedrock Model Distillation trains a small student model to answer like a big teacher for a narrow task. Here is how the job runs, which model pairs are allowed, and why Provisioned Throughput, not the training, decides the cost.
-
Amazon Bedrock Fine-Tuning and Continued Pre-Training (AWS Gen AI Series, Part 16)
When a bigger prompt stops paying off, you change the model itself. A practical walk through fine-tuning and continued pre-training on Amazon Bedrock: which models qualify, what a job costs, and how Nova on-demand hosting changed the math.
-
Amazon Bedrock Prompt Management, Flows, and Prompt Caching (AWS Gen AI Series, Part 15)
Prompt caching, Prompt Management, and Bedrock Flows get grouped together and confused constantly. What each one does, what caching actually saves, and which to reach for.
-
Amazon Bedrock Guardrails, Content Filters, and Grounding Checks (AWS Gen AI Series, Part 14)
Amazon Bedrock Guardrails inspects text into and out of a model across six policies. Where each fits, how to call it inline and standalone, what it costs, and where it trips you in production.
-
Amazon Bedrock Agents, Action Groups, and the AgentCore Shift (AWS Gen AI Series, Part 13)
How Bedrock agents turn one question into several model calls, how action groups and return of control work, what a request really costs, and why new builds now start on AgentCore.
-
Amazon Bedrock Knowledge Bases and Managed RAG, End to End (AWS Gen AI Series, Part 12)
How Amazon Bedrock Knowledge Bases turn your documents into retrieval augmented generation, which vector store and chunking to pick, and why the bill is dominated by a vector-store floor, not tokens.
-
Amazon Bedrock Converse API vs InvokeModel, and When to Use Each (AWS Gen AI Series, Part 11)
InvokeModel hands you each model’s native JSON. The Converse API gives one request and one response shape for every chat model on Bedrock. Here is when to use each, and where InvokeModel is still the only door.
-
Amazon Bedrock Data Residency, KMS, and Security (AWS Gen AI Series, Part 10)
What AWS can and cannot see in a Bedrock call, where your prompts live at rest, and when a customer managed key is worth the operational weight. A practical walk through residency, KMS, IAM, and invocation logging.
-
Amazon Bedrock Private Access with PrivateLink and VPC Endpoints (AWS Gen AI Series, Part 9)
Bedrock traffic leaves your VPC by default. Here is how I wire it shut with PrivateLink interface endpoints, private DNS, and a scoped endpoint policy, plus what it costs per month.
-
Amazon Bedrock Regions, Quotas, and Cross-Region Inference (AWS Gen AI Series, Part 8)
Regions decide which models you can call and how much throughput you get. Here is how Bedrock quotas, token burndown, and cross-Region inference profiles fit together, and how to size a quota request that gets approved.
-
AWS Trainium and Inferentia vs GPUs, and When Each Wins (AWS Gen AI Series, Part 7)
Trainium and Inferentia are Amazon’s own AI chips, not GPUs. Here is when they beat H100 instances on cost, what the Neuron SDK actually demands, and when to stay on GPU.
Architect’s Toolkit
PJ’s Tools
VMware Cloud Foundation
- VCF Documentation
- VCF 9 Planning & Preparation Workbook
- VCF Bill of Materials (BoM)
- VMware Compatibility Guide
- VMware Interoperability Matrix
- VMware Configuration Maximums
- VMware Ports & Protocols
- VMware Hands-on Labs
- RVTools Download
Nutanix
AI & Cloud-Native Platform
- NVIDIA Build (Model Catalog)
- NVIDIA AI Enterprise Reference Architecture
- NVIDIA NIM Performance Benchmarking
- NVIDIA NGC Catalog
- NeMo Microservices Helm Chart
- Helm Charts Repository
- Hugging Face Models
Architecture & Design
About the Author
Dr Pranay Jha
Dr. Pranay Jha is a Cloud and AI Consultant with 18+ years of experience in hybrid cloud, virtualization, and enterprise infrastructure transformation. He specializes in VMware technologies, multi-cloud strategy, and Generative AI solutions. He holds a PhD in Computer Applications with research focused on Cloud and AI, has published multiple research papers, and has been a VMware vExpert since 2016 and a VMUG Community Leader.
You May Have Missed
DrJha