Practical SDLC for AI-Powered Applications
Building an AI-powered application requires more than adding a model to a traditional software project. Teams must manage code, prompts, data, models, and
Building an AI-powered application requires more than adding a model to a traditional software project. Teams must manage code, prompts, data, models, and
The simplest route is a full-stack Next.js application. It runs on Node.js while keeping the pages, authentication, API routes, and styling
A strong way to use Codex for this is to make AGENTS.md define the architecture and engineering rules, then give Codex small issue-
The key is to think of vLLM GPU memory as a budget shared by several very different kinds of data. Some memory is essentially
A useful way to understand an LLM is to separate it into four pieces: architecture + weights + tokenizer + runtime configuration. An inference engine such as
1. Purpose This SOP describes how to use LiveBench to evaluate model quality through a production-style vLLM OpenAI-compatible API. Environment: ComponentConfigurationInference EnginevLLMAPIOpenAI-
1. Purpose This SOP describes a secure and production-ready deployment of vLLM on Amazon Linux 2023, using an NVIDIA GPU-based EC2 G6
Version: 1.0 Platform: Amazon Linux 2023 Deployment Method: Docker (Recommended) Audience: System Administrators, DevOps Engineers, AI/ML Engineers 1. Purpose This document describes
AWS EC2 GPU Instance Comparison FeatureAmazon EC2 G5Amazon EC2 G6Amazon EC2 G7Processor2nd-generation AMD EPYC processorsNewer-generation processors optimized for GPU workloadsCustom Intel Xeon
Document Version: 1.0 Applies To: Ubuntu 22.04/24.04, Debian 12, RHEL/Rocky 9 (or equivalent Linux distributions) Model: google/gemma-4-
1. Purpose This SOP describes how to install and configure Open WebUI on a Linux server using Docker Compose. The deployment supports: * Connection to
1. Purpose This SOP provides instructions for installing and configuring Ollama on a Linux server, enabling remote access, verifying the service, and downloading an
Below is a typical enterprise employee onboarding and offboarding process covering HR, IT, Security, Facilities, Finance, Legal, and the employee’s manager. 1. Employee
1. Purpose This guide defines a minimal but effective IT management process for a startup company developing AI solutions. The goal is to establish
Version 1.0 Date 2026-06-22 Domain mptwork.com Model Identity rooted in the existing Microsoft Entra tenant; hierarchy of Management Groups → Subscriptions
Version 1.0 Date 2026-06-22 Domain mptwork.com Model GCP Organization rooted in Cloud Identity / Google Workspace for mptwork.com; hierarchy of
Version 1.0 Date 2026-06-22 Domain mptwork.com Model AWS Organizations multi-account (management account + member accounts), workforce access via IAM Identity
Version 1.0 Date 2026-06-22 Domain mptwork.com Products Jira + Confluence (Atlassian Cloud), managed in the Atlassian organization at admin.atlassian.com
Version 1.0 Date 2026-06-22 Domain mptwork.com Products Jira + Confluence (Atlassian Cloud), managed in the Atlassian organization at admin.atlassian.com
Version 1.0 Date 2026-06-22 Domain mptwork.com Deployment GitHub Enterprise Cloud with an enterprise account mptwork (github.com/enterprises/mptwork) Identity
Version 1.0 Date 2026-06-22 Domain mptwork.com Deployment GitLab.com (SaaS) with a top-level group mptwork (gitlab.com/mptwork) Identity
Version 1.0 Date 2026-06-22 Tenant context mptwork.com · Sophos Central console central.sophos.com Product Sophos Endpoint (powered by Intercept X)
Version 1.0 Date 2026-06-22 Employee Sally Wong → swong@mptwork.com Device Windows 11 laptop (corporate, provisioned via Windows Autopilot + Microsoft Entra
Version 1.0 Date 2026-06-22 Employee Ben Scully → bscully@mptwork.com Device MacBook Pro (corporate, enrolled via Apple Business Manager + Automated Device
Version 1.0 Date 2026-06-22 Tenant context mptwork.com · Jamf Cloud mptwork.jamfcloud.com Product Jamf Pro (full Apple device management) Companion
Version 1.0 Date 2026-06-22 Example domain mptwork.com Admin console https://admin.google.com Replace the example values (mptwork.com, sample
Version 1.0 Date 2026-06-22 Tenant context mptwork.com (mptwork.onmicrosoft.com) Purpose Create the three organization accounts that Intune needs to
Version 1.0 Date 2026-06-22 Tenant context mptwork.com (mptwork.onmicrosoft.com) Console Microsoft Entra admin center — https://entra.microsoft.com Companion
Version 1.0 Date 2026-06-22 Tenant context mptwork.com (mptwork.onmicrosoft.com) Console Microsoft Intune admin center — https://intune.microsoft.com Companion
Version 1.0 Date 2026-06-22 Example domain mptwork.com Example tenant mptwork.onmicrosoft.com Companion SOP Device management & endpoint protection — sop-
Version 1.1 Date 2026-06-22 Scope Tenant signup → Intune device management (Windows, macOS, iOS, Android) → Microsoft Defender endpoint protection → SOC 2 / ISO
Version 1.0 Date 2026-06-22 Scope 3 subsidiaries (US, Japan, UK) · AWS, GCP, Azure · on-premises offices Companion diagram global-network-hub-
Gateway UniFi UDM Pro Max (US HQ office) VPN client pool 10.12.70.0/24 (the VPN VLAN) Auth backend UDM built-in
On-prem side US HQ office, UniFi UDM Pro Max — 10.12.0.0/16 AWS side US AWS prod VPC, us-west-2
Subsidiary US (10.0.0.0/12) Location US HQ office (on-premises) Gateway UniFi Dream Machine Pro Max (UDM-Pro-Max) On-prem
Subsidiary US (10.0.0.0/12) Provider / region Azure · West US 2 (westus2) Environment prod VNet Regional (vnet-us-prod-westus2), address space
Subsidiary US (10.0.0.0/12) Provider / region GCP · us-west1 (Oregon) Environment prod VPC Global, custom-mode (us-gcp-prod-vpc) Region
Subsidiary US (10.0.0.0/12) Provider / region AWS · us-west-2 (Oregon) Environment prod VPC CIDR 10.1.0.0/20 (4,
Version 1.0 Date 2026-06-22 Status Reference design Scope 3 subsidiaries (US, Japan, UK) · AWS, GCP, Azure · on-premises datacenters Supernet 10.
No playbooks in this category yet.