---
title: "Visual Explainers"
description: "Visual introductions to the engineering concepts behind the tools we use."
---

# Visual Explainers
> Visual introductions to the engineering concepts behind the tools we use.

## Visual Explainers

### [The Car Seat Safety Stack: Four Layers Before the Ride](https://yomu.fyi/visual-explainers/the-car-seat-safety-stack-four-layers-before-the-ride.md)
- Type: Architecture
- Tags: [Architecture](https://yomu.fyi/topic/architecture.md), [Car Seat](https://yomu.fyi/topic/car-seat.md), [Human-in-the-Loop](https://yomu.fyi/topic/human-in-the-loop.md), [Safety](https://yomu.fyi/topic/safety.md), [Workflows](https://yomu.fyi/topic/workflows.md)

A layered architecture for a car-seat ride service: eligibility gates, seat hardware, driver installation training, and caregiver verification. See where checks catch problems before a child rides.

### [How California's SB 53 Turns AI Safety Into a Disclosure Flow](https://yomu.fyi/visual-explainers/how-california-s-sb-53-turns-ai-safety-into-a-disclosure-flow.md)
- Type: How it works
- Tags: [AI Governance](https://yomu.fyi/topic/ai-governance.md), [AI Safety](https://yomu.fyi/topic/ai-safety.md), [California](https://yomu.fyi/topic/california.md), [Compliance](https://yomu.fyi/topic/compliance.md), [Responsible Scaling Policy](https://yomu.fyi/topic/responsible-scaling-policy.md), [Transparency](https://yomu.fyi/topic/transparency.md)

California's SB 53 is a trust-but-verify rule for frontier AI: large developers disclose rather than follow technical mandates. This flow maps the compute threshold, safety frameworks, transparency reports, and the 15-day incident clock.

### [Audience Intelligence: Four Layers from Raw Signals to Decisions](https://yomu.fyi/visual-explainers/audience-intelligence-four-layers-from-raw-signals-to-decisions.md)
- Type: Architecture
- Tags: [Analytics](https://yomu.fyi/topic/analytics.md), [Audience Intelligence](https://yomu.fyi/topic/audience-intelligence.md), [CDP](https://yomu.fyi/topic/cdp.md), [Data Activation](https://yomu.fyi/topic/data-activation.md), [Observability](https://yomu.fyi/topic/observability.md), [Social Listening](https://yomu.fyi/topic/social-listening.md)

Audience Intelligence turns scattered social and behavioral signals into decisions about who to reach and when. This architecture map shows the four layers, what each owns, and where the trade-offs land.

### [Bytecode Interpreter: The Five Parts Between Source and Execution](https://yomu.fyi/visual-explainers/bytecode-interpreter-the-five-parts-between-source-and-execution.md)
- Type: What is it?
- Tags: [Bytecode Interpreter](https://yomu.fyi/topic/bytecode-interpreter.md), [Compilation](https://yomu.fyi/topic/compilation.md), [Compiler](https://yomu.fyi/topic/compiler.md), [MLIR](https://yomu.fyi/topic/mlir.md), [Performance](https://yomu.fyi/topic/performance.md), [WebAssembly](https://yomu.fyi/topic/webassembly.md)

A bytecode interpreter is a compiler's output plus a loop that runs it. This explainer maps its five parts — bytecode, dispatch loop, runtime state, and the host boundary — and why the shape repeats.

### [From Prompt to Approved Write: How Atlas Agent Engine Works](https://yomu.fyi/visual-explainers/from-prompt-to-approved-write-how-atlas-agent-engine-works.md)
- Type: How it works
- Tags: [Agent Memory](https://yomu.fyi/topic/agent-memory.md), [Agents](https://yomu.fyi/topic/agents.md), [Atlas Agent Engine](https://yomu.fyi/topic/atlas-agent-engine.md), [Guardrails](https://yomu.fyi/topic/guardrails.md), [Human-in-the-Loop](https://yomu.fyi/topic/human-in-the-loop.md), [MCP](https://yomu.fyi/topic/mcp.md)

Atlas Agent Engine is the managed runtime that hosts MongoDB agents: it gives each one its own Atlas identity, read-only tools, and an approval gate before any write. Here is the loop, step by step.

### [Anatomy of an Adversarial Prefix Test](https://yomu.fyi/visual-explainers/anatomy-of-an-adversarial-prefix-test.md)
- Type: Architecture
- Tags: [Activation Steering](https://yomu.fyi/topic/activation-steering.md), [Adversarial Prefixes](https://yomu.fyi/topic/adversarial-prefixes.md), [AI Governance](https://yomu.fyi/topic/ai-governance.md), [AI Safety](https://yomu.fyi/topic/ai-safety.md), [Evaluation](https://yomu.fyi/topic/evaluation.md), [Trust Calibration](https://yomu.fyi/topic/trust-calibration.md)

An adversarial prefix is a challenge prepended to a prompt to see whether a model holds its original decision. This diagram maps the layers that turn that pressure test into a measurable, reviewable signal.

### [How to Turn Brand Into a Governed Signal Pipeline](https://yomu.fyi/visual-explainers/how-to-turn-brand-into-a-governed-signal-pipeline.md)
- Type: How to
- Tags: [Advertising](https://yomu.fyi/topic/advertising.md), [Brand](https://yomu.fyi/topic/brand.md), [Branding](https://yomu.fyi/topic/branding.md), [Data Governance](https://yomu.fyi/topic/data-governance.md), [Jobs-to-be-Done](https://yomu.fyi/topic/jobs-to-be-done.md), [Marketing](https://yomu.fyi/topic/marketing.md)

A practical how-to for engineers who need to turn brand from a style guide into a governed signal pipeline: anchor on jobs, model brand roles, unify signals, tokenize meaning, and close the loop with measurement.

### [Branch Like Production: Five Checkpoints Before You Trust It](https://yomu.fyi/visual-explainers/branch-like-production-five-checkpoints-before-you-trust-it.md)
- Type: How to
- Tags: [Branching](https://yomu.fyi/topic/branching.md), [CI/CD](https://yomu.fyi/topic/ci-cd.md), [Database Branching](https://yomu.fyi/topic/database-branching.md), [Data Pipelines](https://yomu.fyi/topic/data-pipelines.md), [Migrations](https://yomu.fyi/topic/migration.md), [Postgres](https://yomu.fyi/topic/postgres.md)

A five-checkpoint workflow for spinning up a database branch that behaves like production: clone config, replay migrations, deploy code, pull fresh data, then verify before you trust it.

### [How to Make Banking Data Agent-Ready: Five Checkpoints](https://yomu.fyi/visual-explainers/how-to-make-banking-data-agent-ready-five-checkpoints.md)
- Type: How to
- Tags: [Banking](https://yomu.fyi/topic/banking.md), [BIAN](https://yomu.fyi/topic/bian.md), [Core Banking](https://yomu.fyi/topic/core-banking.md), [Fintech](https://yomu.fyi/topic/fintech.md), [LTAP](https://yomu.fyi/topic/ltap.md), [Payments](https://yomu.fyi/topic/payments.md)

A five-step path from fragmented payment silos to a canonical, governed record that AI agents can act on: model the payment, unify the copy, stream events, wrap writes in transactions, and scope agent authority.

### [Verified Is Not Trusted: A Five-Step Bot Accountability Loop](https://yomu.fyi/visual-explainers/verified-is-not-trusted-a-five-step-bot-accountability-loop.md)
- Type: How to
- Tags: [Accountability](https://yomu.fyi/topic/accountability.md), [Bot Management](https://yomu.fyi/topic/bot-management.md), [Bots](https://yomu.fyi/topic/bots.md), [Crawlers](https://yomu.fyi/topic/crawlers.md), [Rate Limiting](https://yomu.fyi/topic/rate-limiting.md), [Trust](https://yomu.fyi/topic/trust.md)

A five-step loop for treating bots as accountable actors: verify identity, honor opt-out preferences, measure sustained behavior, band trust, and bound blast radius.

### [Block Devices Explained: Hand the Kernel Chunks, Not Files](https://yomu.fyi/visual-explainers/block-devices-explained-hand-the-kernel-chunks-not-files.md)
- Type: What is it?
- Tags: [Block Device](https://yomu.fyi/topic/block-device.md), [Filesystems](https://yomu.fyi/topic/filesystems.md), [Performance](https://yomu.fyi/topic/performance.md), [virtio-fs](https://yomu.fyi/topic/virtio-fs.md), [Virtual Filesystem](https://yomu.fyi/topic/virtual-filesystem.md), [Virtualization](https://yomu.fyi/topic/virtualization.md)

A block device exposes storage as fixed-size addressable chunks and leaves filesystem logic to the kernel. Here is how that split decides latency, correctness, and who interprets each request.

### [BCVK: A Local Test Bench for Bootable OS Images](https://yomu.fyi/visual-explainers/bcvk-a-local-test-bench-for-bootable-os-images.md)
- Type: Architecture
- Tags: [BCVK](https://yomu.fyi/topic/bcvk.md), [bootc](https://yomu.fyi/topic/bootc.md), [Containers](https://yomu.fyi/topic/containers.md), [Podman](https://yomu.fyi/topic/podman.md), [Testing](https://yomu.fyi/topic/testing.md), [Virtualization](https://yomu.fyi/topic/virtualization.md)

BCVK is a local test bench for bootable container OS images. See how one OCI artifact moves through ephemeral, persistent, and disk-image modes before it reaches real hosts.

### [How to Tune Batch Size: Separate Physical from Logical](https://yomu.fyi/visual-explainers/how-to-tune-batch-size-separate-physical-from-logical.md)
- Type: How to
- Tags: [Batch Size](https://yomu.fyi/topic/batch-size.md), [GPU](https://yomu.fyi/topic/gpu.md), [Hyperparameter Tuning](https://yomu.fyi/topic/hyperparameter-tuning.md), [Inference](https://yomu.fyi/topic/inference.md), [Model Serving](https://yomu.fyi/topic/model-serving.md), [Quantization](https://yomu.fyi/topic/quantization.md)

A practical workflow for choosing batch size in training and inference. Learn to separate physical memory limits from logical update size, then tune with gradient accumulation, quantization, and throughput monitoring.

### [The Basecamp Tradeoff: Small Team, Long Product Life](https://yomu.fyi/visual-explainers/the-basecamp-tradeoff-small-team-long-product-life.md)
- Type: Trade-offs
- Tags: [Basecamp](https://yomu.fyi/topic/basecamp.md), [Mental Models](https://yomu.fyi/topic/mental-models.md), [Product Strategy](https://yomu.fyi/topic/product-strategy.md), [Run Less Software](https://yomu.fyi/topic/run-less-software.md), [Technical Debt](https://yomu.fyi/topic/technical-debt.md)

Basecamp keeps a small team, rewrites the product every few years, and supports old versions indefinitely. This diagram maps what that constraint buys, what it costs, and when it fits.

### [B2B Integration Architecture: Four Layers Between Your Service and the Customer's Stack](https://yomu.fyi/visual-explainers/b2b-integration-architecture-four-layers-between-your-service-and-the-customer-s-stack.md)
- Type: Architecture
- Tags: [B2B](https://yomu.fyi/topic/b2b.md), [Idempotency](https://yomu.fyi/topic/idempotency.md), [Integrations](https://yomu.fyi/topic/integrations.md), [Multi-tenant](https://yomu.fyi/topic/multi-tenant.md), [Observability](https://yomu.fyi/topic/observability.md)

In B2B, most of the engineering lives in the seams: the tenancy boundary, the integration surface, state sync, and the cross-cutting trust rail. This diagram maps those four layers end to end.

### [AWS Resilience Hub: The Trade-offs of Codifying Resilience](https://yomu.fyi/visual-explainers/aws-resilience-hub-the-trade-offs-of-codifying-resilience.md)
- Type: Trade-offs
- Tags: [AWS](https://yomu.fyi/topic/aws.md), [AWS Resilience Hub](https://yomu.fyi/topic/aws-resilience-hub.md), [Compliance](https://yomu.fyi/topic/compliance.md), [Disaster Recovery](https://yomu.fyi/topic/disaster-recovery.md), [Reliability](https://yomu.fyi/topic/reliability.md), [Well-Architected](https://yomu.fyi/topic/well-architected.md)

AWS Resilience Hub turns resilience goals into scoped, shareable policies. This explainer maps three trade-offs: label precision versus upkeep, central policy versus local adoption, and assessment versus drift.

### [Alerting Is a Pipeline, Not a Threshold](https://yomu.fyi/visual-explainers/alerting-is-a-pipeline-not-a-threshold.md)
- Type: Architecture
- Tags: [Alerting](https://yomu.fyi/topic/alerting.md), [Anomaly Detection](https://yomu.fyi/topic/anomaly-detection.md), [Incident Response](https://yomu.fyi/topic/incident-response.md), [Metrics](https://yomu.fyi/topic/metrics.md), [Monitoring](https://yomu.fyi/topic/monitoring.md), [Observability](https://yomu.fyi/topic/observability.md)

Alerting is an architecture, not a threshold: signals flow through detection, sharded evaluation, grouping, and routing. Here is how each layer decides whether a page is real.

### [An Air Gap Is a Pipeline, Not a Wall](https://yomu.fyi/visual-explainers/an-air-gap-is-a-pipeline-not-a-wall.md)
- Type: How it works
- Tags: [Air-Gapped](https://yomu.fyi/topic/air-gapped.md), [Artifacts](https://yomu.fyi/topic/artifacts.md), [Containers](https://yomu.fyi/topic/containers.md), [Local AI](https://yomu.fyi/topic/local-ai.md), [Offline Computing](https://yomu.fyi/topic/offline.md), [Supply Chain](https://yomu.fyi/topic/supply-chain.md)

An air gap is a one-way artifact pipeline, not just a blocked network. This explainer follows six stages, from mirroring dependencies outside to operating inside with no egress.

### [Automated Testing Is a Verification Layer, Not a Test Count](https://yomu.fyi/visual-explainers/automated-testing-is-a-verification-layer-not-a-test-count.md)
- Type: What is it?
- Tags: [Automated testing](https://yomu.fyi/topic/automated-testing.md), [Automation](https://yomu.fyi/topic/automation.md), [CI/CD](https://yomu.fyi/topic/ci-cd.md), [Coding Agents](https://yomu.fyi/topic/coding-agents.md), [Evaluation](https://yomu.fyi/topic/evaluation.md), [Testing](https://yomu.fyi/topic/testing.md)

Automated testing is a verification layer with four moving parts: generate, execute, verify, and triage. This explainer maps each part and shows where human judgment still belongs.

### [Coverage vs. Precision: AI4Bharat's Corpus Pipeline and When to Fine-Tune Instead](https://yomu.fyi/visual-explainers/coverage-vs-precision-ai4bharat-s-corpus-pipeline-and-when-to-fine-tune-instead.md)
- Type: Comparison
- Tags: [AI4Bharat](https://yomu.fyi/topic/ai4bharat.md), [Evaluation](https://yomu.fyi/topic/evaluation.md), [Fine-Tuning](https://yomu.fyi/topic/fine-tuning.md), [Machine Translation](https://yomu.fyi/topic/machine-translation.md), [Multilingual](https://yomu.fyi/topic/multilingual.md), [Synthetic Data](https://yomu.fyi/topic/synthetic-data.md)

Compare two ways to adapt a model to a new language or domain: generate a broad synthetic corpus like AI4Bharat's, or fine-tune on one narrow task and evaluate it.

[Older explainers](https://yomu.fyi/visual-explainers/page/2.md)