| Owner | Muhammad Awais |
| Role | AI Engineer Β· Managing Director @ AI GenMat |
| Education | BS Artificial Intelligence β FAST-NUCES Peshawar |
| Repo | github.com/ik-awais/Python_Projects |
| Level | Scope |
|---|---|
| π’ Beginner | syntax, OOP, data structures, file I/O |
| π‘ Intermediate | automation, data science, APIs, scripting |
| π΄ Advanced | ML, DL, CV, NLP, transformers, diffusion |
| π£ Expert | LLMs, RAG, fine-tuning, agentic systems |
| Domain | Libraries |
|---|---|
| core | Python 3.x Β· OOP Β· decorators Β· generators |
| data | NumPy Β· Pandas Β· Polars Β· SciPy |
| ml | scikit-learn Β· XGBoost Β· LightGBM Β· CatBoost |
| dl | PyTorch Β· TensorFlow Β· Keras Β· JAX |
| cv | OpenCV Β· Pillow Β· Albumentations Β· YOLO |
| nlp | NLTK Β· spaCy Β· Gensim Β· TextBlob |
| llm | HuggingFace Β· LangChain Β· LlamaIndex Β· Ollama |
| agents | LangGraph Β· AutoGen Β· CrewAI Β· Agno |
| rag | FAISS Β· ChromaDB Β· Pinecone Β· Weaviate |
| serve | FastAPI Β· Flask Β· Streamlit Β· Gradio |
| data_eng | Airflow Β· Prefect Β· SQLAlchemy Β· Pydantic |
| viz | Matplotlib Β· Seaborn Β· Plotly Β· Bokeh |
"Every commit is a neuron. Every repo is a brain."
| Folder | Subfolders | Libraries / Notes |
|---|---|---|
| π’ Beginner/ | 01_syntax_basics Β· 02_control_flow Β· 03_data_structures | Core Python fundamentals |
| 04_functions_lambdas Β· 05_oop_classes Β· 06_file_io_os | OOP Β· closures Β· pathlib | |
| 07_error_handling Β· 08_iterators_generators Β· 09_decorators_closures Β· 10_mini_projects | Advanced core + projects | |
| π΅ PAI/ | π data_science/ | NumPy Β· Pandas Β· Polars Β· SciPy |
| π visualization/ | Matplotlib Β· Seaborn Β· Plotly Β· Bokeh | |
| π€ machine_learning/ | scikit-learn Β· XGBoost Β· LightGBM | |
| π§ deep_learning/ | PyTorch Β· TensorFlow Β· Keras Β· JAX | |
| ποΈ computer_vision/ | OpenCV Β· YOLO Β· Pillow Β· Albumentations | |
| π¬ nlp/ | NLTK Β· spaCy Β· Gensim Β· TextBlob | |
| π llm_integration/ | HuggingFace Β· LangChain Β· LlamaIndex | |
| π‘ rag_pipelines/ | FAISS Β· ChromaDB Β· Pinecone Β· Weaviate | |
| πΉοΈ agentic_ai/ | LangGraph Β· AutoGen Β· CrewAI Β· Agno | |
| β‘ serving/ | FastAPI Β· Streamlit Β· Gradio Β· Flask | |
| π οΈ Tools/Files_Garage/ | automation_scripts Β· data_processing Β· utility_helpers | Scripts & helpers |
Every library, framework, and tool used across this repository β organised by domain.
flowchart TD
A(["π Python Core\nsyntax Β· OOP Β· decorators\ngenerators Β· typing"]) --> B
A --> C
B(["π Data Science\nNumPy Β· Pandas Β· Polars\nSciPy Β· Statsmodels"]) --> D
B --> E
C(["βοΈ Scripting & Automation\nos Β· pathlib Β· subprocess\nrequests Β· Typer Β· Rich"]) --> E
D(["π Visualisation\nMatplotlib Β· Seaborn\nPlotly Β· Bokeh Β· Altair"]) --> F
E(["π€ Machine Learning\nscikit-learn Β· XGBoost\nLightGBM Β· CatBoost Β· Optuna"]) --> F
E --> G
F(["π§ Deep Learning\nPyTorch Β· TensorFlow Β· Keras\nJAX Β· Lightning Β· ONNX"]) --> H
F --> I
G(["π¬ NLP\nNLTK Β· spaCy Β· Gensim\nTransformers Β· BERT Β· GPT"]) --> J
H(["ποΈ Computer Vision\nOpenCV Β· YOLO Β· Torchvision\nAlbumentations Β· Detectron2"]) --> J
I(["π LLM Ecosystem\nHuggingFace Β· LangChain\nLlamaIndex Β· Ollama Β· PEFT"]) --> J
I --> K
J(["π‘ RAG Pipelines\nFAISS Β· ChromaDB Β· Qdrant\nPinecone Β· Weaviate"]) --> L
K(["πΉοΈ Agentic AI\nLangGraph Β· AutoGen\nCrewAI Β· Agno Β· Smolagents"]) --> L
L(["β‘ Deployment\nFastAPI Β· Streamlit Β· Gradio\nFlask Β· Pydantic Β· Uvicorn"])
style A fill:#03001e,stroke:#00c8ff,color:#00c8ff
style B fill:#0b0630,stroke:#9d6fff,color:#9d6fff
style C fill:#0b0630,stroke:#9d6fff,color:#9d6fff
style D fill:#1a0845,stroke:#00c8ff,color:#00c8ff
style E fill:#1a0845,stroke:#00c8ff,color:#00c8ff
style F fill:#2a0a6f,stroke:#9d6fff,color:#9d6fff
style G fill:#2a0a6f,stroke:#9d6fff,color:#9d6fff
style H fill:#3a0e8f,stroke:#00c8ff,color:#00c8ff
style I fill:#3a0e8f,stroke:#00c8ff,color:#00c8ff
style J fill:#4a0e8f,stroke:#9d6fff,color:#9d6fff
style K fill:#4a0e8f,stroke:#9d6fff,color:#9d6fff
style L fill:#5a10af,stroke:#00c8ff,color:#00c8ff
π’ Β Beginner Python β click to expand
| # | Topic | Key Concepts |
|---|---|---|
| 01 | Syntax & Basics | variables, types, I/O, operators |
| 02 | Control Flow | if/elif/else, match-case, ternary |
| 03 | Data Structures | list, tuple, dict, set, frozenset |
| 04 | Functions & Lambdas | args, kwargs, closures, functools |
| 05 | OOP | classes, inheritance, dunder methods, dataclasses |
| 06 | File I/O & OS | open(), pathlib, os, shutil, json, csv |
| 07 | Error Handling | try/except, custom exceptions, context managers |
| 08 | Iterators & Generators | iter(), yield, generator expressions |
| 09 | Decorators & Closures | @wraps, @property, @classmethod, @staticmethod |
| 10 | Mini Projects | CLI tools, text games, automation scripts |
π Β Data Science β click to expand
| Library | Purpose | Key APIs |
|---|---|---|
| NumPy | N-dimensional arrays & math | ndarray, broadcasting, linalg, fft |
| Pandas | Tabular data manipulation | DataFrame, groupby, merge, resample |
| Polars | Fast DataFrame (Rust-backed) | LazyFrame, expressions, streaming |
| SciPy | Scientific computing | optimize, stats, signal, sparse |
| Statsmodels | Statistical modelling | OLS, ARIMA, GLM, hypothesis tests |
| Matplotlib | 2D/3D plotting | pyplot, axes, FuncAnimation |
| Seaborn | Statistical visualisation | heatmap, pairplot, FacetGrid |
| Plotly | Interactive charts | graph_objects, express, Dash |
| Bokeh | Web-ready plots | figure, ColumnDataSource, widgets |
π€ Β Machine Learning β click to expand
| Library | Algorithms / Features |
|---|---|
| scikit-learn | LinearRegression, RandomForest, SVM, KMeans, Pipeline, GridSearchCV |
| XGBoost | gradient boosting, GPU support, SHAP integration |
| LightGBM | leaf-wise growth, categorical features, DART |
| CatBoost | ordered boosting, native categoricals, GPU |
| Optuna | TPE sampler, pruning, distributed HPO |
| MLflow | experiment tracking, model registry, serving |
π§ Β Deep Learning β click to expand
| Framework | Highlights |
|---|---|
| PyTorch | nn.Module, autograd, DataLoader, TorchScript, CUDA |
| TensorFlow / Keras | Sequential, Functional API, tf.data, SavedModel |
| JAX | jit, vmap, grad, pmap β functional DL |
| PyTorch Lightning | Trainer, LightningModule, callbacks, auto-logging |
| ONNX | cross-framework model export & runtime |
Architectures covered: CNNs Β· RNNs Β· LSTMs Β· Transformers Β· ViT Β· Diffusion Β· GANs
ποΈ Β Computer Vision β click to expand
| Tool | Use Cases |
|---|---|
| OpenCV | image processing, video I/O, feature detection, homography |
| Ultralytics YOLO | object detection, instance segmentation, pose estimation |
| Pillow | image manipulation, format conversion, filters |
| Albumentations | GPU-accelerated augmentations for training pipelines |
| Torchvision | pretrained models (ResNet, ViT, EfficientNet), transforms |
| Detectron2 | Mask R-CNN, Panoptic FPN, keypoint detection |
π¬ Β NLP β click to expand
| Library | Capabilities |
|---|---|
| NLTK | tokenisation, POS tagging, stemming, WordNet |
| spaCy | NER, dependency parsing, pipelines, custom components |
| Gensim | Word2Vec, Doc2Vec, FastText, LDA topic modelling |
| TextBlob | sentiment analysis, translation, correction |
| HF Transformers | BERT, GPT-2, T5, mT5, RoBERTa, DistilBERT |
| Sentence-Transformers | semantic similarity, bi-encoders, cross-encoders |
π Β LLM Ecosystem β click to expand
| Tool | Role |
|---|---|
| HuggingFace Hub | model & dataset downloads, tokenizers, pipelines |
| PEFT + LoRA | parameter-efficient fine-tuning for large models |
| TRL | RLHF, SFT, DPO, PPO training loops |
| BitsAndBytes | 4-bit / 8-bit quantisation (QLoRA) |
| LangChain | chains, agents, memory, tool use, callbacks |
| LlamaIndex | ingestion pipelines, query engines, node parsers |
| Ollama | local LLM inference (Llama 3, Mistral, Gemma) |
| OpenAI SDK | GPT-4o, embeddings, function calling, vision |
| Anthropic SDK | Claude API, streaming, tool use |
| Groq | ultra-fast inference via LPU hardware |
π‘ Β RAG Pipelines β click to expand
| Component | Tools Used |
|---|---|
| Embedding | text-embedding-3-small, bge-m3, e5-mistral |
| Vector Store | FAISS Β· ChromaDB Β· Pinecone Β· Weaviate Β· Qdrant |
| Retriever | dense, sparse (BM25), hybrid, HyDE |
| Reranker | Cohere Rerank, cross-encoder, FlashRank |
| Generator | GPT-4o Β· Claude Β· Llama 3 Β· Mistral |
| Orchestration | LangChain LCEL Β· LlamaIndex pipeline |
πΉοΈ Β Agentic AI β click to expand
| Framework | Architecture |
|---|---|
| LangGraph | stateful multi-agent graphs, human-in-the-loop |
| AutoGen | conversational multi-agent framework by Microsoft |
| CrewAI | role-based agent crews, task delegation |
| Agno | lightweight agentic framework, tool-use |
| Smolagents | HuggingFace minimal code agents |
Patterns: ReAct Β· Plan-and-Execute Β· Reflexion Β· Tool Use Β· Memory Β· Multi-Agent
| Skill | Level | Progress |
|---|---|---|
| Python Core | Expert | ββββββββββββββββββββ 100% |
| Data Science | Expert | ββββββββββββββββββββ 90% |
| Machine Learning | Advanced | ββββββββββββββββββββ 80% |
| Computer Vision | Advanced | ββββββββββββββββββββ 80% |
| FastAPI / Serving | Advanced | ββββββββββββββββββββ 80% |
| Deep Learning | Advanced | ββββββββββββββββββββ 75% |
| NLP | Advanced | ββββββββββββββββββββ 70% |
| LLM / Fine-tuning | Intermediate+ | ββββββββββββββββββββ 65% |
| RAG Pipelines | Intermediate+ | ββββββββββββββββββββ 65% |
| Agentic AI | Intermediate+ | ββββββββββββββββββββ 60% |
