AI Industry Classification API

NLP + Claude/OpenAI pipeline that classifies companies into industry codes with confidence scoring, human-in-the-loop, and a real-time API on Lambda/SQS.

PythonClaude Sonnet 3.5OpenAINLPVector databasePineconePGVectorAWS LambdaSQSStep FunctionsDynamoDBPostgreSQLAWS RDS

Role

AI Engineer

Timeline

0.7 year

Outcome

Real-time classification API with feedback loop and A/B model comparison.

The story

Automated assignment of NAICS, SIC, and ISIC codes from company names, descriptions, and activities. Custom NLP on Claude Sonnet 3.5 v2 with TF-IDF, embeddings, and NER; hybrid rule-based then ML classification; confidence ranking; human-in-the-loop for low-confidence rows; REST API plus SQS bulk jobs; DynamoDB history; GitHub Actions / CodePipeline / CloudWatch. Role: AI Engineer. Duration: 0.7 year.

Stack

Technologies

AI

PythonClaude Sonnet 3.5OpenAINLPVector databasePineconePGVector

Cloud

AWS LambdaSQSStep FunctionsDynamoDBPostgreSQLAWS RDS