Skip to content
∑ Praveen T N Learning Library
Concepts Flashcards Writing Editorial AI Feed Graph
Portfolio ↗
Overview Concepts Flashcards Writing Editorial AI Feed Graph Search Back to portfolio ↗
Library/Flashcards

Flashcards

11,105 cards in 101 decks, written from the same source as the concepts. Space to flip, arrows to move, K and R to sort what you know from what you do not. Progress lives in your browser.

All decks 01Foundations 02Transformer Internals 03Training & Fine-Tuning 04Reinforcement Learning 05Inference, Systems & Hardware 06Applied LLM Engineering 07Reasoning, Evaluation & Safety 08Multimodal & Applications 09Classical ML & Statistical Learning 10Causal Inference & Experimentation 11Time Series & Forecasting 12Graphs, Recommenders & Structured Data 13Generative Modelling Beyond Transformers 14Efficiency, Compression & Edge AI 15Search & Information Retrieval 16Data & Feature Engineering 17MLOps & Platform Engineering 18Security, Privacy & Adversarial ML 19Governance, Risk & Responsible AI 20Human-AI Interaction, Product & Economics
01

Foundations

The mathematics and neural-network mechanics everything else assumes.

4decks
707cards

Mathematical Foundations

Linear algebra, probability, calculus and optimisation — the machinery every model is built on.

223 cards 16 concepts

Tensors & Neural Plumbing

Shapes, matmuls, forward and backward passes, parameter counts, memory footprints.

211 cards 18 concepts

Deep Learning Building Blocks

Convolutions, recurrence, normalisation, activations, optimisers and regularisation.

112 cards 13 concepts

Information Theory for Language

Entropy, cross-entropy, KL, perplexity, calibration, and language modelling as compression.

161 cards 16 concepts
02

Transformer Internals

Open the box. How a language model actually turns text into predictions.

8decks
929cards

Tokenisation

BPE, WordPiece, Unigram, and the ways subword vocabularies quietly shape model behaviour.

136 cards 15 concepts

Embeddings & Representations

The lookup table, the residual stream, contextual vectors, geometry and superposition.

177 cards 17 concepts

Attention Internals

Queries, keys and values, masking, multi-head and grouped-query, sinks, and the quadratic wall.

94 cards 14 concepts

Positional Encoding

Sinusoidal, learned, RoPE, ALiBi, and how context windows get stretched past training length.

157 cards 16 concepts

Transformer Anatomy

The block, the stack, encoder vs decoder, MoE, and the design choices that separate model families.

58 cards 14 concepts

Training Objectives

Next-token prediction, masked LM, span corruption, fill-in-the-middle, and auxiliary losses.

76 cards 14 concepts

Decoding & Generation

Greedy, beam, temperature, top-k, nucleus, and constrained generation into structured formats.

143 cards 16 concepts

Context & In-Context Learning

Autoregressive generation, the prompt stack, context engineering, and long-context degradation.

88 cards 10 concepts
03

Training & Fine-Tuning

From raw web crawl to an aligned model — data, dynamics, scale and adaptation.

6decks
886cards

Pretraining Data Pipelines

Web-scale corpus construction, filtering, deduplication, decontamination and data mixtures.

140 cards 20 concepts

Synthetic Data

Self-Instruct, distillation, self-play, and how to generate training data without collapsing.

140 cards 20 concepts

Training Dynamics & Scaling

Learning-rate schedules, warmup, loss spikes, critical batch size, muP, and scaling laws.

157 cards 16 concepts

Distributed Training

Data, tensor and pipeline parallelism, ZeRO/FSDP sharding, mixed precision, and offload.

165 cards 13 concepts

Parameter-Efficient Fine-Tuning

LoRA, QLoRA, DoRA, adapters and soft prompts — adapting big models on small budgets.

140 cards 20 concepts

Alignment & Post-Training

SFT, reward modelling, DPO/IPO/KTO/ORPO, model merging, and evaluating an aligned model.

144 cards 21 concepts
04

Reinforcement Learning

Classical RL, then the specific dialect of it that post-trains language models.

5decks
489cards

RL Foundations

MDPs, value functions, TD learning, policy gradients, actor-critic, TRPO and PPO.

140 cards 20 concepts

RL for Language Models

RLHF as an RL problem, KL-regularised objectives, GRPO, RLVR, and reward over-optimisation.

142 cards 21 concepts

Bandits & Exploration

Regret, UCB, Thompson sampling, contextual bandits, and exploration under a budget.

68 cards 7 concepts

Offline & Model-Based RL

Learning from logged data, distribution shift, conservative value estimation, world models and planning.

72 cards 7 concepts

Multi-Agent RL

Self-play, equilibria, credit assignment across agents, emergent coordination and non-stationarity.

67 cards 6 concepts
05

Inference, Systems & Hardware

Where the model meets the silicon, the memory bus and the latency budget.

4decks
645cards

Inference Optimisation

KV cache, FlashAttention, speculative decoding, quantisation and continuous batching.

170 cards 14 concepts

Accelerator Architecture

The memory wall, roofline analysis, GPU execution model, interconnects and systolic arrays.

140 cards 20 concepts

Kernels & Compilers

CUDA, Triton, fusion, tiling, torch.compile, CUDA graphs and roofline-guided optimisation.

140 cards 20 concepts

Serving Systems

Prompt caching, gateways and routing, token accounting, and multi-tenant isolation.

195 cards 13 concepts
06

Applied LLM Engineering

Building things people use: retrieval, prompts, agents and production architecture.

6decks
652cards

Retrieval & RAG

Vector stores, hybrid retrieval and reranking, and when to retrieve instead of fine-tune.

161 cards 14 concepts

Prompt Engineering

In-context learning, chain of thought, structured output, compression and injection-aware design.

94 cards 10 concepts

Agents & Tool Use

Function calling, ReAct loops, MCP, agent memory architectures and evaluation harnesses.

169 cards 15 concepts

Claude Certified Architect

Agentic loops, coordinator-subagent designs, tool interfaces and reliability patterns for Claude.

82 cards 10 concepts

LLM Application Architecture

Routing, fallbacks, caching layers, structured state, and the reference shapes production LLM apps take.

74 cards 7 concepts

AI for Software Engineering

Code models, repository context, patch generation, test-driven agents and SWE benchmarks.

72 cards 7 concepts
07

Reasoning, Evaluation & Safety

Models that think longer, the evals that measure them, and the failure modes that matter.

3decks
748cards

Reasoning Models

Test-time compute, process reward models, the o-series, DeepSeek-R1 and contamination.

196 cards 15 concepts

Evaluation & MLOps

Benchmarks, LLM-as-judge, red-teaming, model registries, drift detection and observability.

144 cards 14 concepts

Safety & Alignment

Prompt injection, jailbreaks, Constitutional AI, reward hacking and mechanistic interpretability.

408 cards 25 concepts
08

Multimodal & Applications

Beyond text — vision, speech, robotics and scientific discovery.

5decks
669cards

Vision & Multimodal

ViT, CLIP, diffusion, SAM, and the vision-language models that read images as tokens.

131 cards 12 concepts

Speech Recognition

Spectrograms, CTC, RNN-T, Conformer, Whisper, streaming, diarisation and self-supervised audio.

140 cards 20 concepts

Speech Synthesis

Acoustic models and vocoders, Tacotron, FastSpeech, HiFi-GAN, neural codecs and voice cloning.

139 cards 20 concepts

Robotics & Embodied AI

Vision-language-action models, action tokenisation, diffusion policies and sim-to-real.

122 cards 10 concepts

AI for Science

AlphaFold, protein language models, materials discovery, and the pitfalls of ML-for-science.

137 cards 12 concepts
09

Classical ML & Statistical Learning

The statistics and non-neural models that still decide most production predictions.

6decks
885cards

Statistical Inference

Estimators, likelihood, the bootstrap, hypothesis testing, multiplicity and what a confidence interval really…

149 cards 7 concepts

Classical Supervised Learning

Linear and logistic regression, regularisation, margins, kernels, and the geometry underneath them.

156 cards 8 concepts

Trees & Ensembles

Impurity splitting, bagging, random forests, gradient boosting, and the engineering inside XGBoost and LightGBM.

145 cards 7 concepts

Unsupervised Learning

Clustering, mixture models and EM, PCA and SVD, manifold embeddings, and density estimation.

145 cards 7 concepts

Bayesian Methods

Priors and posteriors, MCMC and HMC, variational inference, Gaussian processes and model comparison.

145 cards 7 concepts

Feature Engineering

Encoding, missingness, selection, target leakage, and the train-serve skew that eats offline gains.

145 cards 7 concepts
10

Causal Inference & Experimentation

Telling correlation from cause, and measuring whether a change actually helped.

4decks
557cards

Causal Foundations

Potential outcomes, structural causal models, DAGs, confounding, colliders and the do-operator.

137 cards 7 concepts

Online Experimentation

Power, peeking, sample ratio mismatch, variance reduction, interference and long-term effects.

142 cards 7 concepts

Observational Causal Methods

Propensity scores, instrumental variables, difference-in-differences, regression discontinuity and synthetic …

137 cards 7 concepts

Policy Learning & Off-Policy Evaluation

Heterogeneous treatment effects, uplift modelling, importance sampling estimators and doubly robust methods.

141 cards 7 concepts
11

Time Series & Forecasting

Data with an arrow of time, where shuffling the rows destroys the problem.

4decks
377cards

Time Series Foundations

Stationarity, autocorrelation, ARIMA and state space models, seasonality and honest backtesting.

133 cards 7 concepts

Forecasting at Scale

Hierarchical reconciliation, probabilistic forecasts, global models and time-series foundation models.

105 cards 12 concepts

Anomaly & Changepoint Detection

Residual-based detection, CUSUM, seasonal-hybrid methods, and the base-rate problem in alerting.

72 cards 7 concepts

Online & Streaming Learning

Incremental updates, concept drift, regret bounds, and models that must learn from a moving distribution.

67 cards 6 concepts
12

Graphs, Recommenders & Structured Data

Learning over relations, catalogues and columns rather than free text.

4decks
326cards

Graph Neural Networks

Message passing, expressive power and the WL test, over-smoothing, over-squashing and sampling at scale.

74 cards 7 concepts

Knowledge Graphs

Triples and ontologies, entity resolution, embedding-based link prediction, and grounding LLMs in structure.

65 cards 6 concepts

Recommender Systems

Matrix factorisation, two-tower retrieval, ranking objectives, feedback loops and cold start.

120 cards 7 concepts

Tabular Deep Learning

Why trees still win, attention over columns, prior-fitted networks and the benchmarks that decide the argument.

67 cards 6 concepts
13

Generative Modelling Beyond Transformers

Diffusion, flows, adversarial games and state space models, and the theory that connects them.

5decks
388cards

Diffusion Models

Forward noising, denoising objectives, samplers, classifier-free guidance and latent diffusion.

82 cards 7 concepts

Variational & Flow Models

The ELBO, reparameterisation, normalising flows, flow matching and rectified transport.

71 cards 6 concepts

Adversarial Generative Models

The minimax game, mode collapse, Wasserstein critics, and what FID does and does not measure.

76 cards 7 concepts

State Space Models

S4, Mamba, selective scan, the recurrence-convolution duality, and where linear-time sequence models pay off.

94 cards 11 concepts

Energy-Based & Score Models

Unnormalised densities, score matching, Langevin dynamics, and the SDE view that unifies the generative families.

65 cards 6 concepts
14

Efficiency, Compression & Edge AI

Making a model smaller, cheaper and local without giving away the thing that made it useful.

5decks
350cards

Quantisation

Post-training and quantisation-aware methods, outlier channels, GPTQ and AWQ, and low-bit arithmetic formats.

71 cards 6 concepts

Knowledge Distillation

Soft targets and temperature, sequence-level and on-policy distillation, and when a student beats its teacher.

71 cards 6 concepts

Sparsity & Pruning

Magnitude and second-order criteria, structured versus unstructured sparsity, and the hardware that rewards it.

69 cards 6 concepts

Efficient Architectures

Small language models, depth-width tradeoffs, weight sharing, and architectures designed for a latency budget.

65 cards 6 concepts

On-Device & Edge AI

Mobile NPUs, memory-bound inference on consumer silicon, compilation targets and privacy-driven local models.

74 cards 7 concepts
15

Search & Information Retrieval

Thirty years of ranking research that RAG rediscovered, usually the hard way.

5decks
605cards

Classical Information Retrieval

Inverted indexes, TF-IDF and BM25, query processing, and the lexical baselines that refuse to be beaten.

141 cards 7 concepts

Learning to Rank

Pointwise, pairwise and listwise objectives, LambdaMART, position bias and counterfactual training.

133 cards 7 concepts

Query Understanding

Intent classification, spelling and segmentation, expansion, rewriting, and conversational query resolution.

133 cards 7 concepts

Search Evaluation

Pooling and judgments, nDCG and MRR, interleaving, online metrics, and why offline gains vanish online.

129 cards 7 concepts

Vector Databases

Index families, filtered search, freshness and deletes, sharding, and the operational reality of billion-scale ANN.

69 cards 6 concepts
16

Data & Feature Engineering

The pipelines, formats and contracts that decide whether a model ever sees correct inputs.

5decks
381cards

Data Modelling & Storage

Columnar formats, table formats and the lakehouse, partitioning, and modelling choices that decide query cost.

95 cards 11 concepts

Batch & Streaming Pipelines

Event time versus processing time, watermarks, exactly-once semantics, backfills and orchestration.

74 cards 7 concepts

Feature Stores

Offline-online parity, point-in-time correctness, materialisation, and the failure they exist to prevent.

71 cards 6 concepts

Data Quality & Contracts

Expectations and assertions, schema evolution, producer-consumer contracts, and detecting silent corruption.

69 cards 6 concepts

Data Governance & Lineage

Catalogues, column-level lineage, retention and deletion, access control and provenance for training corpora.

72 cards 7 concepts
17

MLOps & Platform Engineering

Everything between a notebook that works and a system that keeps working.

6decks
419cards

Experiment Tracking & Reproducibility

Run metadata, seeds and determinism, environment capture, and what it takes to rebuild a result a year later.

67 cards 6 concepts

Model Registry & Versioning

Artefact lineage, promotion gates, rollback, and versioning models and their data together.

67 cards 6 concepts

CI/CD for ML

Testing pipelines that emit models, data-dependent tests, progressive delivery and automated retraining.

69 cards 6 concepts

ML Observability & Drift

Feature and prediction monitoring, delayed labels, drift statistics, and alerting that does not cry wolf.

71 cards 6 concepts

GPU Fleet & Capacity

Scheduling and quota, fragmentation, preemption, multi-tenancy, and planning capacity under lumpy demand.

76 cards 7 concepts

Cost & FinOps for AI

Unit economics per request, token accounting, reserved versus spot capacity, and attributing spend to features.

69 cards 6 concepts
18

Security, Privacy & Adversarial ML

Attacks on models, data and the supply chain, and the defences that survive contact.

5decks
349cards

Adversarial Robustness

Perturbation attacks, adversarial training, certified defences, and the robustness-accuracy tradeoff.

65 cards 6 concepts

Privacy-Preserving ML

Differential privacy accounting, federated learning, secure aggregation, and the utility cost of each guarantee.

76 cards 7 concepts

LLM Application Security

Injection across trust boundaries, tool and sandbox escape, secret exposure and threat modelling for agents.

74 cards 7 concepts

Model Provenance & Watermarking

Output watermarking, content credentials, fingerprinting weights and detecting extraction.

65 cards 6 concepts

ML Supply Chain Security

Untrusted weights and datasets, deserialisation risk, dependency and registry attacks, and signing artefacts.

69 cards 6 concepts
19

Governance, Risk & Responsible AI

Frameworks, regulation and audit evidence, treated as engineering rather than paperwork.

5decks
340cards

AI Governance Frameworks

NIST AI RMF, ISO/IEC 42001, internal review boards, and turning principles into gates that actually block.

68 cards 7 concepts

AI Regulation & Compliance

The EU AI Act risk tiers, sectoral rules, transparency obligations and evidence a regulator will accept.

66 cards 7 concepts

Fairness & Bias

Group and individual criteria, impossibility results, measurement under missing attributes, and mitigation costs.

67 cards 6 concepts

Transparency & Documentation

Model and system cards, datasheets, disclosure of evaluations, and documentation that survives an audit.

74 cards 7 concepts

AI Assurance & Audit

Third-party evaluation, red-team evidence, incident reporting, and control testing for AI systems.

65 cards 6 concepts
20

Human-AI Interaction, Product & Economics

The people using the system, the product decisions around it, and what compute actually costs.

6decks
403cards

Interaction Design for AI

Latency and streaming affordances, error recovery, steering controls, and designing for probabilistic output.

65 cards 6 concepts

AI Product Management

Scoping around uncertainty, quality bars, offline-to-online metric ladders and shipping under model drift.

65 cards 6 concepts

Trust Calibration & Reliance

Over-reliance and automation bias, confidence display, verification cost and human-AI complementarity.

71 cards 6 concepts

Human Data & Annotation

Guideline design, inter-annotator agreement, preference collection, rater sourcing and label noise.

74 cards 7 concepts

Compute Economics

Capex versus tokens, utilisation and depreciation, price-performance curves and the cost floor of inference.

63 cards 6 concepts

AI Diffusion & Labour

Adoption measurement, task-level exposure, productivity studies and what the evidence does and does not show.

65 cards 6 concepts
The library

1015 concepts, 11,105 flashcards and 123 long-form pieces on AI, NLP, deep learning, LLMs and agentic systems. Free, no sign-up, no paywall.

Sections Concepts Flashcards Writing AI Feed Knowledge Graph
Elsewhere Portfolio Architecture Practice RSS LinkedIn Buy me a coffee

Written and maintained by Praveen T N.

© 2026 Praveen T N