Mastery: connect the pieces

A complete explanation of Training vs inference has four parts. First: Training changes model parameters by learning from examples; inference holds those parameters fixed and uses them to produce predictions. Supplying a prompt or retrieved document during inference changes temporary context, not the model’s learned weights. Second, the mechanism: During training, examples pass forward through the network, a loss compares predictions with targets, gradients flow backward, and an optimizer updates weights. During inference, only the forward computation is needed, often with autoregressive token generation and a KV cache. Fine-tuning is additional training; RAG and prompting are inference-time techniques. Third, the operational controls: training data and split; learning rate; batch size; epochs; optimizer; regularization; checkpoint choice; inference batch size; precision; context length; decoding strategy; and serving hardware. Fourth, the evidence: Training uses loss curves, validation score, generalization gap, data quality, and checkpoint comparisons. Inference uses task quality, calibration, latency, throughput, memory, cost per request, and reliability under realistic load.

Use the scenario as an oral exam: A classifier trains for five epochs on 80,000 labeled tickets and freezes a checkpoint. Every new ticket then runs a forward pass in milliseconds. Adding today’s policy to a prompt helps that request but does not alter the checkpoint. Defend one design choice, then argue against it using this tradeoff: Training is expensive but amortized across many uses and can change persistent behavior. Inference is repeated per request and dominates operating cost at scale. Fine-tuning can improve stable task behavior but is slower to update than retrieval; retrieval gives fresh facts without changing weights. Finally, identify which of these failures your design catches and which remain: claiming prompts retrain the model, leaking evaluation data into training, training on unlicensed or sensitive data, using too high a learning rate, serving a checkpoint with a different tokenizer, ignoring inference drift from prompt changes, and comparing training GPU hours with per-request latency as if they were the same metric.

Mastery is not recalling every term. It is predicting consequences before running the system, noticing when evidence contradicts the prediction, and revising the design without moving the goalposts. Keep a decision record containing the workload, baseline, configuration, test set version, results, known limitations, owner, and rollback condition.

A classifier trains for five epochs on 80,000 labeled tickets and freezes a checkpoint. Every new ticket then runs a forward pass in milliseconds. Adding today’s policy to a prompt helps that request but does not alter the checkpoint.

The worked number is one epoch over 80,000 examples with batch size 100 requires 800 optimizer steps; five epochs require 4,000 steps. State the unit and denominator whenever you report it. A percentage without a denominator can conceal a tiny sample; a latency without a percentile can conceal slow users; a similarity score without a labeled task can conceal irrelevant neighbors. Compare the observed value with a threshold chosen before seeing the final test result.

Now test the tempting shortcut. Suppose the team optimizes only the most visible metric. The result may look better while the system becomes less trustworthy. The reason is concrete: Training is expensive but amortized across many uses and can change persistent behavior. Inference is repeated per request and dominates operating cost at scale. Fine-tuning can improve stable task behavior but is slower to update than retrieval; retrieval gives fresh facts without changing weights. This is why the decision record must include both the intended gain and the tolerated regression. If the tolerated regression is unknown, the change is not ready for a consequential workflow.

• Prefer a measured baseline over a persuasive demo.
• Keep versions, inputs, and thresholds reproducible.
• Separate syntactic success from semantic correctness and authorization.
• Escalate or abstain when evidence falls outside the contract.
• Re-evaluate when data, traffic, models, providers, or user goals change.

These rules turn the topic into an engineering decision rather than a slogan. They also make disagreement productive: another person can challenge the assumptions, rerun the evaluation, and reach a documented conclusion.

Give a three-minute teach-back with no slides. Minute one: define the technique and its boundary. Minute two: trace the mechanism using the worked case and calculation. Minute three: defend the chosen controls with evaluation evidence, then name the strongest unresolved failure. Ask the listener to change one assumption and update your recommendation aloud. You have mastered the topic when the recommendation changes for a technical reason—not because the vocabulary changed—and when you can specify the next experiment that would reduce the most consequential uncertainty.