Nexxiot - IoT Analytics Platform
Python microservices monorepo for AI-powered analytics and data intelligence features for an IoT logistics platform, plus Kubernetes infrastructure configuration.
Key Highlights
- ▹ Built LangGraph-based multi-agent AI system with supervisor routing and SQL generation
- ▹ Designed gRPC-based microservices with async communication and Protocol Buffer schemas
- ▹ Implemented multi-issuer OIDC/JWT authentication with tenant-based access control
- ▹ Architected Alembic migration system with normalized schemas for multi-tenant data isolation
- ▹ Enhanced service scaffolding tooling to auto-generate protobuf definitions and Dockerfiles
Tech Stack
Overview
Nexxiot is an IoT logistics platform where I developed AI-powered analytics features using a Python microservices architecture. The system provides intelligent data analysis capabilities for logistics operations.
Technical Highlights
Multi-Agent AI System
Built LangGraph-based agent orchestration system with supervisor routing, SQL generation, and report generation capabilities using AWS Bedrock (Claude 3.5 Sonnet).
Microservices Architecture
Designed and implemented gRPC-based microservices with async communication, Protocol Buffer schemas, and gateway API layer.
Conversation Management
Built complete conversation persistence system with multi-turn AI interaction support, message history tracking, and feedback collection.
Authentication System
Implemented multi-issuer OIDC/JWT authentication with tenant-based access control and RBAC.
Database Design
Architected Alembic migration system, created normalized schemas with proper indexing for multi-tenant data isolation using CockroachDB.
Developer Tooling
Enhanced service scaffolding tooling to auto-generate protobuf definitions, Dockerfiles, and gateway routes.