# aesedeu/mlops-course architecture

> Architecture diagram of the GitHub repository aesedeu/mlops-course, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-10-04). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/aesedeu/mlops-course
- Repository: https://github.com/aesedeu/mlops-course

## Overview

This repository is a Russian-language MLOps/LLMOps course: learners progress through six lecture-and-workshop topics, with homework covering batch inference, real-time chat/RAG, and real-time NLP/CV. The most concrete runtime examples live in `03_compose_mlops`: Airflow DAGs demonstrate training and inference with MLflow, database access, and S3 file processing; a separate Celery demo illustrates queued tasks. Edges among those runtime components follow the excerpts. Other topic and homework capabilities are documented in the root README; their implementation details are not shown.

Main entry points: [`03_compose_mlops/celery_demo/tasks.py`](https://github.com/aesedeu/mlops-course/blob/2026/03_compose_mlops/celery_demo/tasks.py) (Celery tasks), [`03_compose_mlops/dags/train_dag.py`](https://github.com/aesedeu/mlops-course/blob/2026/03_compose_mlops/dags/train_dag.py) (Wine model training), [`03_compose_mlops/dags/inference_dag.py`](https://github.com/aesedeu/mlops-course/blob/2026/03_compose_mlops/dags/inference_dag.py) (Batch inference), [`03_compose_mlops/dags/db_demo_dag.py`](https://github.com/aesedeu/mlops-course/blob/2026/03_compose_mlops/dags/db_demo_dag.py) (Postgres workflow)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_course["Course topics"]
  node_tools["ML project tools"]
  node_airflow["Airflow"]
end

subgraph group_compose["Compose MLOps examples"]
  node_compose_topic["Compose project"]
  node_training["Wine model training<br/>[train_dag.py]"]
  node_batch_inference["Batch inference<br/>[inference_dag.py]"]
  node_database_demo["Postgres workflow<br/>[db_demo_dag.py]"]
  node_file_ingestion["S3 file workflow<br/>[sensor_demo_dag.py]"]
  node_celery_client["Task client<br/>[client.py]"]
  node_celery_tasks["Celery tasks<br/>[tasks.py]"]
  node_mlflow[("MLflow tracking and registry")]
  node_postgres[("Postgres app database")]
  node_minio[("MinIO object storage")]
  node_redis[("Redis broker and results")]
end

subgraph group_serving["Serving and LLMOps"]
  node_optimized_serving["Triton, TensorRT, OpenVINO"]
  node_llm_serving["vLLM and SGLang"]
  node_llm_gateway["LLM gateway and observability"]
end

subgraph group_homework["Homework"]
  node_batch_hw["Batch inference assignment"]
  node_rag_hw["Chat and RAG assignment"]
  node_realtime_hw["NLP and CV assignment"]
end

node_learner(("Learner"))

node_learner -->|"studies"| node_tools
node_learner -->|"studies"| node_airflow
node_learner -->|"studies"| node_compose_topic
node_learner -->|"studies"| node_optimized_serving
node_learner -->|"studies"| node_llm_serving
node_learner -->|"studies"| node_llm_gateway
node_learner -->|"completes"| node_batch_hw
node_learner -->|"completes"| node_rag_hw
node_learner -->|"completes"| node_realtime_hw
node_training -->|"logs and registers"| node_mlflow
node_batch_inference -->|"loads and logs"| node_mlflow
node_database_demo -->|"reads and writes"| node_postgres
node_file_ingestion -->|"waits and archives"| node_minio
node_celery_client -->|"submits tasks"| node_celery_tasks
node_celery_tasks -->|"configures broker and backend"| node_redis

click node_tools "https://github.com/aesedeu/mlops-course/tree/2026/01_tools"
click node_airflow "https://github.com/aesedeu/mlops-course/tree/2026/02_airflow"
click node_compose_topic "https://github.com/aesedeu/mlops-course/tree/2026/03_compose_mlops"
click node_training "https://github.com/aesedeu/mlops-course/blob/2026/03_compose_mlops/dags/train_dag.py"
click node_batch_inference "https://github.com/aesedeu/mlops-course/blob/2026/03_compose_mlops/dags/inference_dag.py"
click node_database_demo "https://github.com/aesedeu/mlops-course/blob/2026/03_compose_mlops/dags/db_demo_dag.py"
click node_file_ingestion "https://github.com/aesedeu/mlops-course/blob/2026/03_compose_mlops/dags/sensor_demo_dag.py"
click node_celery_client "https://github.com/aesedeu/mlops-course/blob/2026/03_compose_mlops/celery_demo/client.py"
click node_celery_tasks "https://github.com/aesedeu/mlops-course/blob/2026/03_compose_mlops/celery_demo/tasks.py"
click node_optimized_serving "https://github.com/aesedeu/mlops-course/tree/2026/04_triton_tensorrt_openvino"
click node_llm_serving "https://github.com/aesedeu/mlops-course/tree/2026/05_vllm_sglang"
click node_llm_gateway "https://github.com/aesedeu/mlops-course/tree/2026/06_llm_gateway"
click node_batch_hw "https://github.com/aesedeu/mlops-course/tree/2026/hw/01_batch_inference"
click node_rag_hw "https://github.com/aesedeu/mlops-course/tree/2026/hw/02_realtime_chat_rag"
click node_realtime_hw "https://github.com/aesedeu/mlops-course/tree/2026/hw/03_realtime_nlp_cv"

classDef toneNeutral fill:#f8fafc,stroke:#334155,stroke-width:1.5px,color:#0f172a
classDef toneBlue fill:#dbeafe,stroke:#2563eb,stroke-width:1.5px,color:#172554
classDef toneAmber fill:#fef3c7,stroke:#d97706,stroke-width:1.5px,color:#78350f
classDef toneMint fill:#dcfce7,stroke:#16a34a,stroke-width:1.5px,color:#14532d
classDef toneRose fill:#ffe4e6,stroke:#e11d48,stroke-width:1.5px,color:#881337
classDef toneIndigo fill:#e0e7ff,stroke:#4f46e5,stroke-width:1.5px,color:#312e81
classDef toneTeal fill:#ccfbf1,stroke:#0f766e,stroke-width:1.5px,color:#134e4a
class node_tools,node_airflow toneBlue
class node_compose_topic,node_training,node_batch_inference,node_database_demo,node_file_ingestion,node_celery_client,node_celery_tasks,node_mlflow,node_postgres,node_minio,node_redis toneAmber
class node_optimized_serving,node_llm_serving,node_llm_gateway toneMint
class node_batch_hw,node_rag_hw,node_realtime_hw toneRose
class node_learner toneIndigo
```

## Components

### Course topics

- [ML project tools](https://github.com/aesedeu/mlops-course/tree/2026/01_tools): `01_tools`
- [Airflow](https://github.com/aesedeu/mlops-course/tree/2026/02_airflow): `02_airflow`

### Compose MLOps examples

- [Compose project](https://github.com/aesedeu/mlops-course/tree/2026/03_compose_mlops): `03_compose_mlops`
- [Wine model training](https://github.com/aesedeu/mlops-course/blob/2026/03_compose_mlops/dags/train_dag.py): `03_compose_mlops/dags/train_dag.py`
- [Batch inference](https://github.com/aesedeu/mlops-course/blob/2026/03_compose_mlops/dags/inference_dag.py): `03_compose_mlops/dags/inference_dag.py`
- [Postgres workflow](https://github.com/aesedeu/mlops-course/blob/2026/03_compose_mlops/dags/db_demo_dag.py): `03_compose_mlops/dags/db_demo_dag.py`
- [S3 file workflow](https://github.com/aesedeu/mlops-course/blob/2026/03_compose_mlops/dags/sensor_demo_dag.py): `03_compose_mlops/dags/sensor_demo_dag.py`
- [Task client](https://github.com/aesedeu/mlops-course/blob/2026/03_compose_mlops/celery_demo/client.py): `03_compose_mlops/celery_demo/client.py`
- [Celery tasks](https://github.com/aesedeu/mlops-course/blob/2026/03_compose_mlops/celery_demo/tasks.py): `03_compose_mlops/celery_demo/tasks.py`
- **MLflow tracking and registry**
- **Postgres app database**
- **MinIO object storage**
- **Redis broker and results**

### Serving and LLMOps

- [Triton, TensorRT, OpenVINO](https://github.com/aesedeu/mlops-course/tree/2026/04_triton_tensorrt_openvino): `04_triton_tensorrt_openvino`
- [vLLM and SGLang](https://github.com/aesedeu/mlops-course/tree/2026/05_vllm_sglang): `05_vllm_sglang`
- [LLM gateway and observability](https://github.com/aesedeu/mlops-course/tree/2026/06_llm_gateway): `06_llm_gateway`

### Homework

- [Batch inference assignment](https://github.com/aesedeu/mlops-course/tree/2026/hw/01_batch_inference): `hw/01_batch_inference`
- [Chat and RAG assignment](https://github.com/aesedeu/mlops-course/tree/2026/hw/02_realtime_chat_rag): `hw/02_realtime_chat_rag`
- [NLP and CV assignment](https://github.com/aesedeu/mlops-course/tree/2026/hw/03_realtime_nlp_cv): `hw/03_realtime_nlp_cv`

### Other components

- **Learner**

## Connections

- Learner → ML project tools: studies (evidence: `README.md`)
- Learner → Airflow: studies (evidence: `README.md`)
- Learner → Compose project: studies (evidence: `README.md`)
- Learner → Triton, TensorRT, OpenVINO: studies (evidence: `README.md`)
- Learner → vLLM and SGLang: studies (evidence: `README.md`)
- Learner → LLM gateway and observability: studies (evidence: `README.md`)
- Learner → Batch inference assignment: completes (evidence: `README.md`)
- Learner → Chat and RAG assignment: completes (evidence: `README.md`)
- Learner → NLP and CV assignment: completes (evidence: `README.md`)
- Wine model training → MLflow tracking and registry: logs and registers (evidence: `03_compose_mlops/dags/train_dag.py`)
- Batch inference → MLflow tracking and registry: loads and logs (evidence: `03_compose_mlops/dags/inference_dag.py`)
- Postgres workflow → Postgres app database: reads and writes (evidence: `03_compose_mlops/dags/db_demo_dag.py`)
- S3 file workflow → MinIO object storage: waits and archives (evidence: `03_compose_mlops/dags/sensor_demo_dag.py`)
- Task client → Celery tasks: submits tasks (evidence: `03_compose_mlops/celery_demo/client.py`)
- Celery tasks → Redis broker and results: configures broker and backend (evidence: `03_compose_mlops/celery_demo/tasks.py`)
