Intermediate concepts in AI & ML, page 11
Reflection: Teaching LLM Agents to Learn from Mistakes
Reflection gives an agent an "inner monologue" to learn from its mistakes. An Actor model attempts a task, an Evaluator scores it, and a Self-Reflection model generates linguistic feedback for the next try.
AI Transparency: Explaining the Black Box's 'Why'
AI transparency means seeing the 'why' behind an algorithm's decision, not just its code. It's vital for high-stakes systems like credit scoring or news feeds. The footgun is thinking open-sourcing the model is enough; true transparency explains the logic.
TensorFlow Serving: A Production Server for ML Models
Think of TensorFlow Serving as a dedicated web server for your ML models. It provides a stable API for inference and manages model versions, abstracting away deployment complexity. The main footgun is thinking it only serves models; it serves any 'Servable'.
LLMs as Tool Makers: Write Once, Solve Many
LLMs can create their own tools, not just use them. A powerful model writes a reusable function once, and a cheaper model calls it many times. This gives top-tier results at a lower cost for repetitive tasks.
AI Accountability: Who's Responsible When AI Fails?
AI accountability means someone is answerable for an AI's actions. It requires organizations to manage risks and trace decisions throughout the AI's lifecycle, ensuring systems function properly and align with human-centric values.
MLOps: When to Build vs. Buy Your Infrastructure
Deciding to build or buy MLOps tools hinges on whether it creates a competitive advantage. For commodity tasks like experiment tracking, buying a managed service avoids locking up engineers.
Full Fine-Tuning: Updating Every Model Parameter
Full fine-tuning updates all weights of a pre-trained model on your new data, unlike methods that only change a small fraction. Use it to deeply embed new knowledge, but beware: it's costly and risks making the model forget its original general skills.
Fairness Metrics: Auditing Your AI for Bias
Fairness metrics are statistical checks to see if your model's decisions are biased against certain groups. They're crucial for automated systems in hiring or loan approvals.
Model Registry: Source of Truth for Deployed Models
A model registry is the source of truth for which trained model runs where, turning anonymous artifact files into versioned, staged assets. It matters when you deploy multiple models or need instant rollbacks.
Modality Gap: When Multimodal LLMs Don't Trust Their Senses
A multimodal LLM has a modality gap when it trusts one input type (like text) over another (like images), even with identical information. This bias causes performance drops, like ignoring visual data if conflicting text is present.
Datasheets for Datasets: The Nutrition Label for Data
A datasheet is like a nutrition label for a dataset, documenting its origins, contents, and intended use. This is crucial for high-stakes ML systems where hidden biases could cause harm.
Log Transformation: Compress the Long Tail
Log transformation compresses the long tail of skewed data so outliers cannot dominate loss. Use it for features like income or latency that span orders of magnitude. The footgun is blindly applying it to zeros or negatives, which destroys data.
Model Pruning: Making LLMs Smaller, Not Dumber
Model pruning is surgical weight loss for an LLM, removing neurons or layers to reduce its size. It's used to create smaller, faster versions of models like LLaMA for efficient deployment. The footgun: naive pruning can cripple the model's core capabilities.
AI Safety: Preventing Unintended Consequences
AI Safety is the engineering discipline for preventing intelligent systems from causing harm, by accident or misuse. It's crucial for autonomous systems like self-driving cars or large models that can amplify bias.
Zero-Shot Learning: Classifying the Unseen
Zero-Shot Learning lets a model classify things it never trained on. It works by linking visual features to semantic descriptions, like identifying a 'zebra' from the description 'striped horse'. The footgun is assuming it creates knowledge from nothing.
Experiment Run: The Immutable Training Receipt
An experiment run is an auto-generated log for one training job: it captures hyperparameters, metrics, code, and artifacts. Teams use runs to debug regressions and audit settings. The footgun is logging many metrics without versioning data so comparison fails.
Dynamic Batching: Balancing LLM Throughput and Latency
Dynamic batching groups LLM requests like a bus that leaves on a schedule or when full, whichever comes first. This improves throughput in inference servers by avoiding long waits. The footgun: all requests in a batch are still held hostage by the slowest one.
Prototypical Networks: Learning from a Handful of Examples
Prototypical Networks classify new categories from few examples by finding the average representation, or 'prototype,' for each class. This is key for few-shot image recognition where you have only 1-5 examples.
MLflow Models Standardize Deployment Packaging
MLflow Models wrap artifacts into a standard package so one pipeline serves sklearn or PyTorch without new deployment code. Teams ship experiments to REST endpoints without Dockerfiles per model. Missing dependency logging lets model load but fail to predict.
Quantization-Aware Training (QAT): Forcing Models to Learn While Quantized
QAT forces a model to "learn its own compression" by simulating quantization during training. This lets you shrink LLMs to aggressive low-bit formats (like 4-bit) where simpler post-training methods fail.
We are hiring for this. Every open role lists the topics its interview covers, so you can prepare for the real thing rather than guessing.
See open roles