Davy Costa

I’m a Software Engineer/Data Analyst who builds end-to-end data and AI solutions that ship reliably in production. I’ve architected scalable data lake infrastructure on AWS and delivered ETL/ELT pipelines that turn raw data into analytics-ready datasets—using Medallion (Bronze/Silver/Gold) for clear, performant processing.

On the data engineering side, I implemented Terraform-based infrastructure, built AWS Glue + PySpark ETL/ELT workflows, and optimized storage for analytical queries in S3 as Parquet (Athena). I also developed serverless ingestion with AWS Lambda in Python for automated batch processing, orchestrated dependencies and data quality checks with AWS Glue, and supported ML training workflows using SageMaker and MLflow—then deployed models with SageMaker Endpoints, ECR (Docker), and AWS Step Functions for inference and re-training orchestration.

More recently, I led GenAI solutions with RAG on GCP, using LangChain, LangGraph, and ADK—accelerating decision-making for financial clients by 30%. I built multi-agent orchestration for faster responses (50%), added security and Guardrails to reduce hallucinations, optimized token usage and latency through chunking/re-ranking/caching, and created an internal GenAI integration framework in Python (CLI/SDK) with FastAPI and Pydantic adopted by multiple teams—cutting integration time by 40% while leveraging FAISS, ChromaDB, and Pinecone for semantic retrieval.