Portfolio

A selection of production-focused AI and data engineering projects, including LLM evaluation frameworks, RAG systems, Snowflake/dbt pipelines, and multi-cloud AI architecture. Full source on GitHub.

Multi-Cloud Serverless RAG

Problem

The local RAG pipeline was tied to a single machine and one AI provider — no way to compare AWS, Azure, and GCP AI stacks on the same workload.

Architecture
Terraform
AWS — Glue + OpenSearch Serverless + Lambda + Bedrock Azure — Azure ML + AI Search + Functions + AI Foundry GCP — Vertex AI + Firestore + Cloud Functions
Hugging Face Spaces
Technologies
Python Terraform AWS Bedrock Azure AI Foundry Vertex AI OpenSearch Serverless Azure AI Search Firestore Streamlit
Result

One RAG system deployable on any major cloud with a single terraform apply — live on Hugging Face Spaces with one page per cloud backend.

Multi-Cloud Serverless RAG screenshot click to enlarge
Why it matters

Enterprise AI teams rarely operate on a single cloud — they inherit existing infrastructure, face vendor lock-in decisions, or need to compare AI stack costs across providers. Seeing the same pipeline built three ways makes those trade-offs concrete.

RAG Pipeline

Problem

500 arXiv research papers were unsearchable via keyword search — a paper on 'reducing compute for LLMs' never surfaces when you search 'efficient LLM training', even though it's exactly what you need.

Architecture
arXiv API PDF parser + chunker sentence-transformers embedder PostgreSQL + pgvector Claude Sonnet 4.6 generator Streamlit chat UI
Technologies
Python PostgreSQL pgvector sentence-transformers Claude API Streamlit Docker LangChain
Result

A chat interface that answers natural-language questions grounded in the paper collection, with cited sources — deployed locally via Docker.

Why it matters

LLMs answer from training data that can be outdated, hallucinated, or simply wrong for your domain. RAG grounds every answer in your own documents, making responses verifiable and traceable to the source.