AI/ML Services Portfolio

6 major AI/ML services delivered at AWS scale, serving millions of users

🔍

Knowledge Base

Multi-format content processing pipeline that ingests AWS training materials from PDF, xAPI, HTML, PowerPoint, Markdown, and CloudFormation templates. Uses advanced NLP to parse and vectorize content, enabling semantic and lexical search capabilities through OpenSearch with vector embeddings.

Impact

6 Content Formats
100K+ Documents Indexed
Vector Embeddings NLP OpenSearch Semantic Search
⚖️

Evaluation Service for AI-Generated Content

Automated quality assurance system that validates AI-generated AWS training content against official AWS documentation. Uses LLM-as-a-judge architecture to evaluate correctness, accuracy, and alignment with authoritative sources. Accelerates development of reliable AI-generated content while maintaining high quality standards for the AWS Training Business.

Impact

10x Faster Content Validation
95%+ Accuracy Rate
LLM as Judge Content Validation Quality Assurance Bedrock
🧠

Knowledge Graph

Intelligent content modularization system that transforms AWS documentation and digital content into reusable learning units with mapped relationships and dependencies. Leverages Graph database to maintain knowledge graph , while storing learning objectives and content modularization metadata as vector embeddings in OpenSearch for semantic and lexical search. Leverages MCP (Model Context Protocol) to enable AI agents to dynamically generate and package personalized learning paths on-demand based on specific learning goals.

Impact

50K+ Modular Learning Units
500K+ Knowledge Relationships
Knowledge Graph Neptune Open Search MCP Amazon Strands Bedrock
💡

Data Sense - Intelligent Data Lake Query Service

Agentic AI framework that transforms natural language queries into actionable data insights. Leverages dual knowledge bases (OpenSearch for semantic/lexical search and Neptune graph database for entity relationships) to understand data lake metadata enriched through rigorous LLM evaluation. Query Generation Agent translates user questions into SQL, while Query Execution Agent runs Athena queries and synthesizes results into human-readable insights powered by generative AI.

Impact

10x Faster Data Discovery
Zero SQL Knowledge Required
Agentic AI Natural Language Query OpenSearch Neptune Athena Bedrock
🎯

Digital Recommendations

AI-powered recommendation system that matches learners with relevant AWS training courses . Built using collaborative filtering and content-based algorithms.

Impact

50K+ Daily Recommendations
12% Engagement Lift
Recommendation System Machine Learning Personalization
📊

Multilingual Feedback Analytics & Sentiment Intelligence

Advanced NLP-powered feedback collection and sentiment analysis system that processes customer feedback in multiple languages for AWS Training products and services. Automatically generates actionable insights including Content Satisfaction Score (CSAT) and Net Promoter Score (NPS), enabling data-driven decisions to improve training quality and learner experience at scale.

Impact

15+ Languages Supported
100K+ Monthly Feedback Analyzed
Sentiment Analysis NLP Multilingual CSAT/NPS Amazon Comprehend