Self Service Dashboards enabled with AI chat

Project Technical Requirements
Objective
Develop an integrated data pipeline and analytics platform that enables automated data extraction, transformation, and loading (ELT) from an ERP system to a modern data warehouse, with an interactive dashboard and a natural language chat interface powered by a basic LLM.
________________________________________
1. ELT Pipeline
• Source Database:
• Type: MySQL 5
• Role: Primary ERP data source
• ETL Tool:
• Platform: Airbyte (open source)
• Function: Automated extraction and loading of data from MySQL to the target database
• Target Database:
• Type: PostgreSQL (latest stable version, optionally with TimescaleDB for time-series data)
• Role: Central data warehouse for analytics and transformations
• Transformation:
• Tool: dbt (core version)
• Function: SQL-based transformations to prepare data for analytics and visualization
• Scope: Data cleaning, aggregation, and preparation for dashboard consumption
________________________________________
2. Dashboard and Self-Service Analytics
• Frontend Framework:
• Platform: Next.js (latest version, using App Router)
• Language: TypeScript
• Visualization:
• Libraries: charjs and/or Recharts for interactive charts and graphs
• Features:
• Comparison graphs (time series, segment comparisons, etc.)
• Dynamic filtering and customizable visualizations
• Self-Service:
• Functionality:
• Users can select dimensions and measures
• Drag-and-drop interface for building and arranging charts
• Real-time data exploration with minimal IT support
• Performance:
• Caching: Implement Redis or similar for query result caching
• Rendering: Server-side rendering (SSR) and incremental static regeneration (ISR) for fast load times
• Real-time Updates: WebSocket integration for live data updates
________________________________________
3. LLM Chat Interface
• Model:
• Type: Basic large language model
• Integration: local deployment
• Query Engine:
• Framework: LangChain
• Function: Converts natural language queries into SQL and executes them against PostgreSQL
• Security:
• Validation: SQL parser with allow-list for permitted operations to prevent injection
• Scope: Users can ask questions in plain language and receive visualized answers
• User Experience:
• Interface: Chat-based, integrated into the Next.js dashboard
• Output: Visualized data and summarized insights based on user queries
________________________________________
4. Infrastructure and Deployment
• Deployment:
• Containerization: Docker for all major components (Airbyte, PostgreSQL, Next.js app)
• Orchestration: Docker Compose for local development, Kubernetes for production (optional)
• Monitoring:
• Pipeline: Airbyte monitoring and Grafana dashboards
• Application: New Relic or similar for application and database performance monitoring
• Security:
• Data: TLS encryption for all database connections
• Authentication: NextAuth.js with OAuth2 providers (Azure AD, Google)
• Authorization: Row-level security in PostgreSQL for data access control
• Audit: Database audit triggers and log archiving
________________________________________

Please provide examples of similar work done.