Learning path
Agent Performance
Understand tool-call execution, cache behaviour, retries, and trajectory evaluation in stateful AI agents.
View curriculumAI engineering education and research
Useful intelligence is an engineering problem.
Learn how language models work, optimize local inference, and build reliable AI agents through guided lessons and reproducible experiments.
The interesting work begins after the demo works.
Structured learning
Featured learning paths
Learning path
AI Systems Foundations
Understand the components of an AI application and the boundaries between models, retrieval, tools, and production services.
View curriculumLearning path
Evaluation & Benchmarking
Design measurements that distinguish speed, quality, retrieval effectiveness, and reliable agent execution.
View curriculumLearning path
Local LLM Inference
Run and measure open-weight models locally, connecting quantization, hardware limits, latency, throughput, and output quality.
View curriculumLearning path
Model Internals
Connect model execution, token generation, context, and KV-cache behaviour to memory use and serving performance.
View curriculumLearning path
Production AI Infrastructure
Build reliable AI services through explicit system boundaries, policy, validated outputs, durable execution, evaluation, and workload admission.
View curriculumIndependent publication
Recent engineering writing
Before the Pod Exists: How Kueue Decides Whether an AI Job May Run
Learn how Kueue admits Kubernetes AI and batch jobs, manages scarce GPU quota, and hands workloads to kube-scheduler.
Read noteEvaluation Is the Control System for Production AI - Production AI Systems - Part 5
Evaluation connects product requirements, versioned datasets, component failures, operational constraints, release gates, and production feedback.
Read noteLong-Running AI Work Is a Durable Workflow - Production AI Systems - Part 4
Long-running AI work should be modeled as durable state transitions with idempotent stages, checkpoints, bounded retries, cancellation, and backpressure.
Read noteFrom Model Output to Trusted Application Data - Production AI Systems - Part 3
Schema-constrained generation is only the first trust boundary. Production systems still need evidence, policy, authorization, bounded repair, and explicit disposition.
Read noteThe Model Gateway Is a Policy Boundary, Not a Proxy -Production AI Systems - Part 2
A model gateway should centralize provider compatibility, routing, quotas, and data policy without absorbing product workflow logic.
Read note