Learn/Course 18
Production · weeks 22–23

Serving & inference systems

Launch traffic causes p99 spikes and leaves expensive GPUs idle.

Opening production brief

A measurable problem, before a model.

Users

Operators, product owners, and people affected by the serving systems decision.

Data

prediction contracts, REST/gRPC, batching, timeouts with version, owner, event time, and availability contracts.

Constraints

Launch traffic causes p99 spikes and leaves expensive GPUs idle. Every solution must state latency, cost, capacity, and fallback limits.

Success

Offline evidence plus a deployable resolution: Batch and online paths with batching, autoscaling, load tests, canaries, and rollback.

CHAPTER 1

Frame and baseline

Turn the production problem into explicit data, metric, and baseline contracts.

  1. 01

    Choose batch, online, streaming, edge, or hybrid

    Place inference where the decision’s deadline, freshness, and failure semantics can actually be met.

    65–90 min · checkpoint · deterministic lab
  2. 02

    Package model, preprocessing, schema, and runtime

    The deployable artifact is the complete raw-input-to-output program, not the weights alone.

    65–90 min · checkpoint · deterministic lab
CHAPTER 2

Build and stress

Formalize the machinery, test counterexamples, and expose system limits.

  1. 03

    Features, batching, caching, and backpressure

    Fast serving comes from bounded, version-aware work; graceful overload is a feature, not an afterthought.

    65–90 min · checkpoint · deterministic lab
  2. 04

    Capacity, load testing, autoscaling, and tails

    Design capacity around the tail and the delay of the scaling loop, not around average throughput.

    65–90 min · checkpoint · deterministic lab
CHAPTER 3

Resolve and operate

Join model behavior to architecture, observability, rollout, and rollback.

  1. 05

    Shadow, canary, rollback, fallback, observability

    A release is safe only when exposure is attributable, bounded, observable, and quickly reversible.

    65–90 min · checkpoint · deterministic lab
  2. HW

    Production inference contract

    Specify a typed service with warmup, batching, resilience, load testing, rollout, metrics, and fallback.

    Rubric · staged hints · reference resolution
Prerequisite guidance

3 concepts make this faster.

Prerequisites are guidance, never hard gates.

C12GPU trainingA profiled, mixed-precision, memory-budgeted training loop.C16Efficient inferenceA capacity plan using efficient attention, quantization, batching, and modern routing.C17ML platformsA traced batch/stream platform with orchestration, registry, lineage, and ownership.
Ready to work the problem?

Choose batch, online, streaming, edge, or hybrid

Start lesson 01