Raman Gaur

Architected and operationalized a multi-agent automation workflow using LangGraph and OpenAI/Anthropic APIs for Goldman Sachs financial operations, automating multi-step document processing, data extraction, validation, and enrichment. Engineered automated data movement pipelines and integrated LangFuse tracing, automated retries, and validation checks, reducing error-handling triage time by 75%.

Led the design of agentic AI document processing platforms using custom ML models and LLMs, streamlining classification, entity extraction, and structured metadata generation across document-heavy enterprise workflows. Integrated AI model outputs with PostgreSQL, Pinecone, and Neo4j databases, and architected cloud-native microservices on AWS ECS using Docker.

Engineered NLP-based data extraction and classification pipelines in Python and C++, processing structured and unstructured operational document streams under rigid performance SLAs. Built RESTful backend microservices on AWS with exponential backoff retries and failure monitoring, and standardized Docker deployment workflows with GitHub Actions.

Developed predictive ML and data processing pipelines in Python and SQL (PostgreSQL/Oracle), executing complex data transformation, validation, and feature engineering routines. Constructed automated data validation rules and transformation scripts, and collaborated with client leads to scope requirements and conduct UAT.

Developed Python and SQL scripts for enterprise data classification, automated imputation, and feature standardization across backend relational databases. Monitored production execution logs, identified software defects, and applied backend hotfixes to ensure reliable data movement during enterprise release cycles.

Bachelor of Technology in Computer Science and Engineering from Galgotias University.