Hey there! 📚 Welcome to TS-01

TS 01 · Study Notes · AI/ML

Every concept explained so simply. Every number computed step-by-step.

---

🏆 Featured Projects — Applied Portfolio

Beyond coursework: end-to-end applied work spanning cybersecurity, deep learning, and MLOps — built, demoed, tested, and documented.

🔓 Reentrancy Attack — Smart Contract Security Live exploit + defense on Ethereum smart contracts — drains an 11 ETH vault via a classic reentrancy bug, then blocks the identical attack with Checks-Effects-Interactions + a nonReentrant guard. Built with Foundry, Anvil, and a Next.js/MetaMask UI; proven with Forge tests exposing the full recursive call trace. CSL6010 Cyber Security · Group 6 🛡️ PoisonedRAG + RAG-Shield — RAG Poisoning Defense Reproduces a USENIX Security 2025 RAG-poisoning attack (~90% attack success from just 5 malicious documents), then builds RAG-Shield — a 3-ring defense-in-depth pipeline (ingest screening, retrieval trust scoring, cross-LLM consensus across Claude/Mistral/LLaMA) that drives attack success down to ~13% while preserving normal-query accuracy. CSL6010 · Group 6

🎨 Mode Collapse in GANs — Presentation Kit Solo-built deep-learning presentation kit: an interactive browser-based live demo that simulates GAN mode collapse in real time (ten target classes collapsing to one, live mode-coverage tracking, one-click fix reveal), a presenter runbook, a Q&A guide, and full study notes running theory → numericals → runnable PyTorch (simple GAN + DCGAN). 📦 Repo · 🎥 Live Demo · Deep Learning · Group 1 🎭 MLOps Emotion Pipeline — DistilBERT CI/CD End-to-end MLOps: fine-tunes distilbert-base-uncased for 6-class emotion classification (Kaggle GPU training + W&B experiment tracking), publishes to a public Hugging Face model repo, packages inference in Docker, and ships via GitHub Actions CI/CD with a branch-protected, peer-reviewed PR workflow. MLOps · Group 12

🎓 More Deep Learning Demos — Batchmates' Work

Interactive live demos from fellow M.Tech AI batchmates, worth a look:

Demo	Topic	Live Link
🎨 Watching a GAN Collapse	Generative Adversarial Networks — mode collapse, live	rpaut03l.github.io/gan-mode-collapse-demo-grp-1-iit-j
🎬 RBM Movie Recommender	Restricted Boltzmann Machines for recommendation	teal-frangipane-2c6927.netlify.app
🧠 DBN Visualization	Deep Belief Networks, visualized	nikhilsaini-iitj.github.io/dbn-visualization
🧲 Contrastive Learning Explorer	Explore how neural networks learn by comparing positive and negative pairs in embedding space	scarlet-hatti-38.tiiny.site
🔗 Contrastive Learning (Group 21)	Self-supervised contrastive learning	g25ait2134-tech.github.io/DL_Contrastive_Learning_Group21
♻️ Transfer Learning — Feature Reuse	Transfer learning & feature reuse	sureshbabugandla.github.io/transfer-learning-feature-reuse

---

🎒 What This Repo Contains

	Subject	Description
🤖 AI	Artificial Intelligence	Search, Logic, Planning, Bayesian Networks, Reinforcement Learning — 21 topics, 9800+ lines of ELI5 notes
🧠 ML	Machine Learning	Ensemble Methods, Boosting, AdaBoost, Gradient Descent, Regularization, model building
🔢 Maths	Mathematics for AI/ML	Linear Algebra, Probability & Statistics, Optimization, Calculus foundations
💻 DSA&T	Data Structures & Techniques	Arrays, Trees, Graphs, Dynamic Programming, Sorting & Searching
📈 ODS	Optimization for Data Science	Convex Optimization, Gradient Descent, Convergence Analysis, Constrained Optimization, Duality

---

⚡ Focus Areas

Area	Topics Covered
🔍 Search & Optimization	BFS · DFS · A* · IDA* · Hill Climbing · Simulated Annealing · Genetic Algorithms
🧩 Constraint Satisfaction	Backtracking · AC-3 · Forward Checking · MRV · LCV · Min-Conflicts
♟️ Game Playing	Minimax · Alpha-Beta Pruning · Expectimax · Evaluation Functions
📐 Logic & Reasoning	Propositional Logic · First-Order Logic · Unification · Resolution · Backward Chaining
🗺️ Planning	Situation Calculus · STRIPS · Partial Order Planning · Frame Problem
🕸️ Probabilistic Models	Bayesian Networks · CPTs · d-Separation · Causality · Simpson's Paradox · do-Calculus
🎮 Reinforcement Learning	MDP · Bellman Equation · Value Iteration · Q-Learning · REINFORCE · Actor-Critic
🧠 ML Algorithms	Ensemble Methods · AdaBoost · Gradient Boosting · Regression · Clustering · Deep Learning
📈 Optimization	Convex Functions · Gradient Descent · SGD · Convergence Rates · Duality · KKT Conditions

---

🌟 What Makes This Special

	Feature	Details
🍼	ELI5 Explanations	Every concept starts with a story — ice cream shops for Queues, plate piles for Stacks, melting ice cream for Discount Factor γ
🧮	Full Arithmetic Traces	No hand-waving. Every f=g+h, every e^(ΔE/T), every P(B|J,M) computed with every multiplication shown
⚠️	Exam Trap Alerts	200+ common mistakes flagged — the ones professors LOVE to test and students ALWAYS get wrong
📊	Worked Examples	100+ step-by-step traces: BFS/DFS on graphs, A* on Romania, AC-3 REVISE calls, Q-table updates
🔗	Linked Resources	Every topic maps to: class data (ts-01), best YouTube lecture, AIMA textbook chapter

---

🗺️ AI — Complete Topic Navigator

🔍 Search (Quiz 1) # Topic Lines 01 Uninformed Search (BFS, DFS, UCS, IDS) 1061 02 Informed Search (Greedy, A*) 645 03 Memory-Bounded (IDA*, RBFS, SMA*) 305 04 Local & Evolutionary Search 544 05 And-Or Search 295 🧩 CSP (Quiz 1 + Assignment) # Topic Lines 06 Backtracking, AC-3, MRV, LCV 539 07 Min-Conflicts 328 ♟️ Adversarial Search (Quiz 1) # Topic Lines 08 Minimax Algorithm 329 09 Alpha-Beta Pruning 332 10 Expectimax Search 407

📐 Knowledge & Reasoning (Quiz 2) # Topic Lines 11 Propositional Logic 435 12 FOL — Syntax & Semantics 376 13 FOL — Inference & Unification 622 🗺️ Planning (Major) # Topic Lines 14 Situation Calculus 249 15 STRIPS & Sub-goals 355 16 Partial Order Planning 229 🕸️ Bayesian & Causality (Quiz 2 + Major) # Topic Lines 17 Bayesian Networks 443 18 Causality & Probabilistic Reasoning 238 🎮 Reinforcement Learning (Major) # Topic Lines 19 MDP & Policy 624 20 Q-Learning, Passive & Active RL 442 21 Policy Search & REINFORCE 538

---

🔗 Continue the Journey — TS-02 (DLOps & MLOps)

TS-01 covers the foundations (AI, ML, Maths, DSA&T, ODS). The applied, production-facing continuation — Deep Learning Operations and ML Operations — lives in the sibling repo TS-02.

🔥 DLOps PyTorch → CNNs → tracking → distributed → deployment

#	Topic	Notebook
01	Intro to PyTorch — tensors, autograd, first nn	.ipynb
02	Basics for DL — activations, losses, optimizers	.ipynb
03	CNN + Feature Extraction — CIFAR/LeNet, RandomForest hybrid	.ipynb
04	Datasets & DataLoaders — transforms, augmentation	.ipynb
05	Custom Datasets — ImageFolder, TinyVGG	.ipynb
06	TensorBoard — SummaryWriter, PR curves, hparams	.ipynb
07	W&B Sweeps (course) — init/log/sweep/agent	.ipynb
08	W&B Sweeps (official) — sweep_config grammar	.ipynb
09	W&B Artifacts — data + model versioning	.ipynb
10	Distributed Training — DataParallel, model parallel	18a · 18b · 18c
11	TorchScript — trace vs script	.ipynb
12	ONNX — export, checker, onnxruntime	.ipynb

📎 Start here → DLOps Hub · README · all notebooks

⚙️ MLOps Data → pipelines → containers → orchestration → production

Area	Covers
Systems Concepts	ML system design, reproducibility, the "why" behind MLOps
Preprocessing & EDA	Data cleaning, feature pipelines, exploratory workflows
Git · Docker · K8s	Containerizing training/serving, versioned pipelines, orchestration
Experiment Tracking	Bridges directly into DLOps modules 06-09 (TensorBoard, W&B)
Deployment	Serving patterns that pair with DLOps modules 11-12 (TorchScript, ONNX)

🔗 Related: K8s MLOps pipeline repo · ML workflows blog

How the three repos fit together: TS-01 (theory/foundations) → TS-02/MLOps (systems & pipelines) → TS-02/DLOps (deep learning in production) — read in that order, or jump straight to whichever layer you need.

---

🛠️ Tech Stack

📚 Study & Notes

🐍 AI / ML / RL

☁️ Cloud & Infra

🔧 MLOps / AIOps

---

📖 Key References

Resource	Type	Link
AIMA (Russell & Norvig)	📕 Textbook	artint.info
Stanford CS221	🎓 Lectures	YouTube
MIT AI (Patrick Winston)	🎓 Lectures	YouTube
IIT Delhi AI	🎓 Lectures	YouTube
Turing — "Can Machines Think?"	📄 Paper	PDF

---

⭐ Star this repo if it helped you study!

Built with ❤️ and ☕ by rpaut03l

AI/ML Enthusiast · Singapore 🇸🇬