AI Agents – EB1A Experts

ai

An AI agent application designed to automate operations for the EB1A Experts team. The system includes content generation, verification using custom lexical rules and NER, and a public-facing AI client application for end users.

System Design & Architecture

Architecture Overview

The system follows a microservices architecture deployed on AWS EKS. It uses Kong API Gateway for authentication and rate limiting. The backend services include AI content generation, content verification using NLP models, and CRUD APIs. Monitoring and logging are handled through Grafana stack (Grafana, Jaeger, Loki).

Key Components

  • Kong API Gateway: Authentication, rate limiting, and routing
  • AI Content Generation Service: OpenAI and LLM providers integration
  • Content Verification Service: Custom lexical rules (NLTK) and NER (SpaCy, BERT)
  • Backend CRUD APIs: FastAPI-based RESTful services
  • Monitoring Stack: Grafana, Jaeger (tracing), Loki (logging)
  • Container Orchestration: AWS EKS (Kubernetes)
  • Security: Trivy scanning, SonarQube code quality

Data Flow

User request → Kong API Gateway (auth/rate limit) → EKS cluster → AI Content Service / Verification Service → LLM providers / NLP models → Response → Monitoring stack → User.

Technologies Used

AWS EKSKong API GatewayFastAPIGrafanaJaegerLokiNLTKSpaCyBERTOpenAIDockerKubernetes

Architecture References

Kong API Gateway

Cloud-native API gateway for authentication, rate limiting, and API management.

Learn More

AWS EKS

Amazon Elastic Kubernetes Service for container orchestration and deployment.

Learn More

Grafana Observability Stack

Grafana for visualization, Jaeger for distributed tracing, and Loki for log aggregation.

Learn More

SpaCy & BERT for NER

Named Entity Recognition using SpaCy and BERT models for content verification.

Key Features

  • AI-powered content generation and verification
  • Custom lexical rules and NER for content validation
  • Automated CI/CD with comprehensive monitoring
  • Kong API Gateway with security features
  • EKS deployment and management

My Responsibilities

  • Built AI agent application to automate content writing for various teams using OpenAI and other LLM providers
  • Developed AI agents to generate and verify content required for EB1A applications
  • Implemented content verification using custom lexical rules with NLTK and named entity recognition with SpaCy and BERT models
  • Automated CI/CD pipeline and deployed application on AWS EC2
  • Integrated logging and monitoring using Grafana, Jaeger, Loki, and related services
  • Developed backend CRUD APIs and implemented AI-driven content generation and verification
  • Setup Kong API Gateway with authentication and rate limiting
  • Handled security tasks including Trivy scan and SonarQube setup
  • Worked on setting up EKS

Technologies Used

GenAIPythonFastAPIAWS ServicesGrafanaJaegerKong API GatewayEKSNLTKSpaCyBERT