samantha

Senior Data Scientist

Sole owner of the ML/AI stack and data infrastructure at a 30-person startup.. Built user-facing AI agents with LangGraph that query company databases through the Model Context Protocol (MCP), helping clients find insights in their data.. Designed a RAG pipeline on AWS Bedrock with a pgvector store and built its evaluation framework: eval dataset, groundedness, cost per query and human review.. Built an LLM invoice classifier on AWS Bedrock over unstructured documents, tuning the model and prompts to improve groundedness.. Served models through FastAPI on AWS (SageMaker, Lambda, S3) for real-time inference and owned them through release and monitoring.. Built PySpark pipelines (ingestion, feature prep, scoring, automated retraining) under Git.. Founded a biweekly AI/ML workshop (60% of the company, 4 teams) and write the documentation that aligns product and engineering.

Data Scientist

Data Analyst

Owned ML DevOps for three production models (deployment, monitoring, maintenance).. Built cohort, survival and churn models that guided retention strategy.. Built SQL and Python ETL pipelines feeding model outputs into near-real-time workflows.

Freelance Data Scientist

Built a neural-network invoice fraud-detection system compliant with Mexican Tax Authority (SAT) requirements.. Built Monte Carlo simulation and forecasting models for logistics and retail optimization clients.