AWS Architecture Showcase
Powered by Amazon Web Services
Chatbot with LangChain and S3
A serverless chatbot architecture using LangChain for processing and S3 for storage.
Components:
- API Gateway: Accepts user input (e.g., chatbot query).
- Lambda: Processes the query and calls LangChain.
- LangChain: Decides what to do (e.g., fetch knowledge, call external APIs).
- S3: Stores chat logs or files generated by the AI.
- DynamoDB: Stores metadata (e.g., user sessions, preferences).
Event-Driven AI Workflow
An event-driven architecture for AI-powered file processing and analysis.
Components:
- S3: Stores uploaded files (e.g., documents for analysis).
- S3 Event: Triggers Lambda when a file is uploaded.
- Lambda: Processes the file and calls LangChain for AI reasoning.
- LangChain: Generates insights and stores results in DynamoDB.
- EventBridge: Publishes an event when processing is complete.
- SQS: Queues tasks for downstream processing.
Asynchronous Task Processing
An asynchronous architecture for processing AI tasks using a queue-based approach.
Components:
- API Gateway: Accepts user requests.
- Lambda: Validates the request and sends a task to SQS.
- SQS: Queues the task for processing.
- Lambda: Polls SQS, processes the task, and calls LangChain.
- LangChain: Executes the AI workflow and stores results in DynamoDB.
Multi-Model AI Pipeline
A comprehensive AI pipeline that handles multiple types of input and uses various AI services.
Components:
- API Gateway: Accepts user input (text, image, or speech).
- Lambda Orchestrator: Directs input to appropriate AI services.
- LangChain: Processes text input and generates final response.
- Amazon Rekognition: Analyzes image input.
- Amazon Transcribe: Converts speech input to text.
- Lambda Aggregator: Combines results from different AI services.
- DynamoDB: Stores processed results and metadata.
Serverless AI Training Pipeline
A serverless architecture for training and deploying AI models.
Components:
- S3: Stores datasets and trained models.
- Lambda: Triggers training jobs and model deployment.
- SageMaker: Runs training jobs and hosts deployed models.
- SNS: Notifies data scientists of job completion.
- API Gateway: Provides inference endpoint for deployed models.
- LangChain: Processes inference requests and responses.