Build an AI system
Give a model context, tools, and a way to be evaluated.
TOPICS / FOLLOW YOUR CURIOSITY
Small concepts.
Connected understanding.
Look inside the systems you use. Start with a question, follow a mechanism, and connect it to the next idea.
A few good places to begin.
Suggested reading trailsTrace a web request
Start with a name. Follow it to an application.
Ship a service
Connect a change to its runtime and operating signals.
THE KNOWLEDGE MAP
Explore by subject.
AI engineeringLLMs · RAG · agents · evaluation53 topics
From tokens and retrieval to agents, evaluation, and production model systems.
- Next-Token Predictionhow autoregressive models generate text word by word~60 sec read
- Context Windowthe working memory limit of a single model request~60 sec read
- Attentionhow models calculate context-sensitive word relationships~75 sec read
- Transformersthe foundational neural network architecture of modern LLMs~80 sec read
- Semantic Searchfinding documents by conceptual meaning rather than exact keywords~65 sec read
- Retrievalfetching external, real-time knowledge at query time~70 sec read
- Context Engineeringcurating and structuring optimal prompt payloads for LLM stability~75 sec read
- Distillationcompressing knowledge from a large teacher model into a smaller student model~70 sec read
- LLMlarge language models that generate text by predicting the next token~85 sec read
- RAGgrounding LLM answers in your own data~80 sec read
- Vector Databasesstoring and searching by semantic similarity, not exact keywords~75 sec read
- Prompt Engineeringcrafting inputs to get reliable, structured outputs from LLMs~70 sec read
- Fine-Tuningspecializing a base model on your domain data~80 sec read
- Agents & Tool UseLLMs that take actions — not just talk~85 sec read
- Model Evaluationmeasuring LLM output quality systematically~75 sec read
- Embeddingsconverting text into numerical vectors for semantic search~75 sec read
- Guardrailspreventing LLMs from producing harmful, unsafe, or off-brand output~80 sec read
- Multimodal ModelsAI models that understand text, images, audio, and video together~80 sec read
- Tokenizationhow LLMs break text into pieces their neural networks can process~75 sec read
- Prompt Cachingoptimizing LLM API latency and token reuse costs~70 sec read
- Structured Outputsenforcement of JSON schemas on LLM response payloads~75 sec read
- Local LLM Inferencerunning open-source models on your developer machine~75 sec read
- Model Routingsending each prompt to the best model for the job~75 sec read
- AI Gatewaysa unified proxy layer between your app and multiple LLM providers~80 sec read
- External Integrationsconnecting AI agents to real-world APIs and SaaS platforms~70 sec read
- Agent Harnessthe orchestration framework that turns an LLM into a reliable agent~80 sec read
- Execution Loopthe perceive → reason → act cycle that drives autonomous agents~70 sec read
- Memory Systemsgiving agents short-term and long-term recall beyond the context window~80 sec read
- Tokens & Contextwhat the model can 'see'~65 sec read
- Inference vs Trainingusing vs building a model~65 sec read
- GPUs & VRAMwhy AI needs special hardware~65 sec read
- Quantizationshrinking models to fit~65 sec read
- Temperature & Samplingwhy answers vary~65 sec read
- Model Servingputting a model behind an endpoint~70 sec read
- MCP & Function Callinghow models talk to your systems~70 sec read
- Evalstesting something non-deterministic~65 sec read
- Latent Spacehigh-dimensional space of meanings~65 sec read
- Parameterslearnable weights in the model~60 sec read
- Pre-trainingtraining on massive unlabeled data~65 sec read
- Base Modelthe foundational pre-trained model~60 sec read
- Instruction Tuningfine-tuning to follow instructions~65 sec read
- Alignmentensuring helpful, honest, and harmless outputs~65 sec read
- RLHFReinforcement Learning from Human Feedback~70 sec read
- System Promptsetting the model's behavior and constraints~65 sec read
- User Promptthe question or task from the end user~60 sec read
- Zero-Shot Promptingasking to perform tasks without examples~60 sec read
- Few-Shot Promptingproviding examples to guide the model~65 sec read
- Chain of Thoughtthinking step-by-step before answering~65 sec read
- Top-k Samplinglimiting choices to the top k tokens~65 sec read
- Top-p Samplingchoosing from cumulative probability p~65 sec read
- Groundinganchoring answers in verified facts~65 sec read
- LoRA & QLoRAfine-tuning models on consumer hardware~65 sec read
- DSPycompiling declarative prompt pipelines~70 sec read
NetworkingDNS · TLS · HTTP · routing17 topics
Follow a request through names, connections, routing, and the web.
- DNShow names become IP addresses~60 sec read
- TLShow encryption happens before the lock icon appears~75 sec read
- HTTPthe protocol the web speaks~65 sec read
- TCP/IPthe foundation all internet communication rides on~70 sec read
- Load Balancersdistributing traffic across servers to keep things fast and reliable~80 sec read
- API Gatewaya single front door for all your backend services~85 sec read
- CDNserving content from a server near your user~75 sec read
- TCP Handshakehow two machines agree to talk~60 sec read
- HTTP Request Lifecyclewhat happens when you hit Enter~75 sec read
- Ports & Socketshow one machine runs many services~60 sec read
- NAThow private IPs reach the internet~65 sec read
- Subnets & CIDRwhat /24 really means~70 sec read
- Routinghow packets find their way~60 sec read
- Reverse Proxythe middleman in front of your app~65 sec read
- Firewallswho gets in, who gets out~60 sec read
- Latency vs Bandwidthwhy 'fast' means two things~60 sec read
- WebSocketskeeping a connection alive~65 sec read
Linux & systemsProcesses · permissions · shell · memory13 topics
Read the operating system through its processes, files, permissions, and services.
- Processes & Signalshow programs live and die~65 sec read
- File Permissionsrwx and who you are~60 sec read
- Filesystem Hierarchywhy /var, /etc, /opt exist~60 sec read
- Pipes & Redirectionhow commands chain~65 sec read
- Environment Variableswhere config hides~60 sec read
- Users & Sudothe permission model~60 sec read
- Systemd & Serviceswhat keeps daemons running~65 sec read
- Cron & Schedulingrunning things later~60 sec read
- Memory & Swapwhat 'out of memory' means~65 sec read
- File Descriptorswhy 'too many open files' happens~70 sec read
- Symlinks & Hard Linkstwo names for one file~65 sec read
- Exit Codeshow scripts know something failed~60 sec read
- Shell Startup Chain.bashrc vs .profile~70 sec read
Containers & KubernetesDocker · Kubernetes · Helm · persistence16 topics
Understand images, workloads, orchestration, and the boundaries around a service.
- Dockerpackage apps with their entire runtime into lightweight boxes~80 sec read
- Kubernetesorchestrating containers at scale~90 sec read
- Container Layershow images are built from stacked, cached snapshots~65 sec read
- Helmkubernetes package manager — templated charts for complex apps~70 sec read
- Docker Composedefine and run multi-container apps with a single YAML file~70 sec read
- Container Registrywhere container images are stored, versioned, and scanned~70 sec read
- Kubernetes Ingressrouting external traffic to services inside your cluster~75 sec read
- Sidecar Patternattaching helper containers to your pod without modifying app code~75 sec read
- Containers vs VMsisolation without a whole OS~70 sec read
- Container Networkinghow containers find each other~65 sec read
- Volumes & Persistencedata that survives restarts~65 sec read
- Orchestrationwhy one container is never enough~70 sec read
- Health Checksalive isn't ready~65 sec read
- Resource Requests & Limitssharing a node politely~65 sec read
- Namespaces & Multi-tenancymany teams, one cluster~70 sec read
- StatefulSets & Operatorsrunning databases in clusters~75 sec read
Cloud architectureIAM · VPC · serverless · storage19 topics
Connect compute, storage, identity, and networks with cost and resilience in mind.
- IAMwho can do what on which resource~80 sec read
- Serverlessrun code without provisioning or managing servers~85 sec read
- Object Storageunlimited, highly durable storage for any blob of data~70 sec read
- VPCyour private network in the cloud~80 sec read
- Auto Scalingautomatically adjusting compute capacity to match demand~80 sec read
- SQSfully managed message queuing for decoupling services~80 sec read
- DynamoDBserverless NoSQL database with single-digit millisecond latency~85 sec read
- Route 53AWS's DNS service — also handles failover and traffic routing~75 sec read
- Node Autoprovisioningjust-in-time Kubernetes node provisioning on AWS~70 sec read
- Regions & AZswhere your stuff physically is~65 sec read
- Block vs Object Storagethree ways to keep bytes~70 sec read
- Managed vs Self-hostedwhat you pay the cloud to worry about~65 sec read
- Cloud Billingwhy the bill surprises everyone~65 sec read
- Shared Responsibilitywhat the cloud won't secure for you~70 sec read
- Multi-region & DRsurviving a region outage~70 sec read
- Egress Costswhy data leaving costs money~65 sec read
- Spot Instancescheap machines that vanish~70 sec read
- AWS PrivateLinkprivate connections without public internet~75 sec read
- Cell-Based Architecturecontaining blast radius through cells~80 sec read
Delivery & operationsCI/CD · observability · rollbacks · SLOs17 topics
Move changes through delivery, observe their behavior, and recover when things fail.
- CI/CDautomated pipelines that build, test, and deploy every change~80 sec read
- Monitoring & Observabilityknowing what your system is doing at all times~90 sec read
- Configuration Managementautomating server setup and software configuration at scale~75 sec read
- Artifact Managementstoring, versioning, and distributing build outputs~60 sec read
- Canary Deploymentsrolling out changes to a small subset before full release~80 sec read
- Feature Flagsturning features on and off without deploying code~75 sec read
- Incident Managementthe structured process of detecting, responding to, and learning from outages~85 sec read
- Service Meshmanaging microservice traffic, security, and observability out-of-band~80 sec read
- GitOps CDdeclarative GitOps continuous delivery for Kubernetes~75 sec read
- Blue-Green Deploymentsshipping without downtime~75 sec read
- Declarative vs Imperativedescribing vs commanding~65 sec read
- Idempotencywhy running twice must be safe~60 sec read
- Semantic Versioningwhat 2.1.3 promises~65 sec read
- Rollbacksundoing a bad release~65 sec read
- SLI / SLO / SLAmeasuring 'is it working?'~70 sec read
- Configuration Driftwhen reality stops matching the code~65 sec read
- OpenTelemetrystandardizing logs, metrics, and traces~70 sec read
Infrastructure as codeTerraform · state · modules · GitOps6 topics
Turn infrastructure into definitions you can review, repeat, and maintain.
- Terraform Statethe source of truth Terraform uses to reconcile real-world resources~85 sec read
- Terraform Modulesreusable, versioned infrastructure components~70 sec read
- GitOpsgit as the single source of truth for deployments~75 sec read
- Infrastructure as Codemanaging infrastructure through machine-readable definition files~65 sec read
- Pulumiinfrastructure as code using real programming languages~80 sec read
- Packerbuilding golden machine images with code~80 sec read
Security & identityOAuth · secrets · least privilege · supply chain22 topics
Reason about trust, access, secrets, and the controls between system boundaries.
- OAuth 2.0how apps get limited access to your data without your password~90 sec read
- Zero Trustnever trust, always verify — no implicit access based on network location~85 sec read
- Secrets Managementstoring and rotating API keys, passwords, and certificates safely~75 sec read
- CORSthe browser's same-origin policiy enforcer~70 sec read
- SIEMcentralized logging and alerting for security events~75 sec read
- JWTcompact, self-contained tokens for passing identity and claims~80 sec read
- WAFfiltering malicious traffic before it reaches your application~75 sec read
- mTLSmutual TLS — both sides prove their identity, not just the server~85 sec read
- SBOMa full inventory of every software component in your application~80 sec read
- OIDC Federationsecretless repository authentication for CI/CD actions~80 sec read
- Network Policiesin-cluster firewall controls for isolating pod traffic~70 sec read
- AuthN vs AuthZwho you are vs what you may do~60 sec read
- Public-key Cryptographythe math behind SSH and TLS~65 sec read
- SSH Keys & Agentspasswordless, but safe~65 sec read
- Hashing vs Encryptionone-way vs two-way~65 sec read
- Certificates & PKIwhy browsers trust anyone~70 sec read
- Least Privilegethe default should be 'no'~65 sec read
- Secrets Rotationkeys that expire on purpose~70 sec read
- Injection Attackswhen input becomes code~65 sec read
- Encryption Rest vs Transittwo places data leaks~65 sec read
- CVEs & Patchinghow vulnerabilities get names~70 sec read
- ZTNAperimeter-less access verification~75 sec read
Git & collaborationCommits · branches · conflicts · reflog8 topics
Understand the history behind commits, branches, merges, and safe recovery.
- Commits & the DAGhistory as a graph~65 sec read
- Branching & Mergingparallel work, one history~65 sec read
- Rebase vs Mergetwo ways to combine work~70 sec read
- The Staging Areawhy git has three states~60 sec read
- Remotes & Trackingyour copy vs everyone's copy~65 sec read
- Merge Conflictswhen two truths collide~65 sec read
- Tags & Releasespinning a moment in history~60 sec read
- The Refloggit's undo button~65 sec read
No matching concepts yet.
Try a shorter phrase, another subject, or clear the filters.