Rafael Madrigal
Senior Machine Learning Engineer @Manulife
Signup · Get unlimited contacts
WORK HISTORY
Senior Machine Learning Engineer @Manulife
Toronto, ON, CA
Architected scalable GenAI systems as part of Global Contact Centre Transformation across USA (John Hancock) and Canada (Manulife). 9 Knowledge Assistants serving users, utilizing a hybrid architecture using LangGraph that combines massive unstructured document collections (Azure AI Search) with large structured customer data (SQLMI) for personalized answer generation• Engineered a hierarchical, two-stage retrieval architecture (Azure AI Search) to mitigate semantic noise across large document collections utilized hybrid search, dynamic OData filtering, and custom Lucene queries and scoring profiles to maximize retrieval precision@k• Authored clean and maintainable Python core libraries and crafted feature-parity roadmaps that standardized the RAG lifecycle (document pre-processing to evaluation); established reusable code and API designs that slashed initial UAT prototyping from 3 months to 1-2 weeks and total time-to-production from 7 months to 3 months• Used Microsoft Azure tools including AI Search, OpenAI, Databricks, Foundry. Built Agentic Workflows using Langchain and LangGraph.
EDUCATION
Asian Institute of Management
Master of Science, Data Science, with High Distinction
ABOUT RAFAEL MADRIGAL
Senior ML & Research Engineer with 10 years of experience – last 5 years in the end-to-end translation of SOTA AI Research into production systems within regulated finance. My background combines the experience as an ML Research investigating algorithmic feasibility with an ML Engineer architecting enterprise-scale solutions. This dual perspective enables me to lead the full research-to-deployment lifecycle: from the exploration of Privacy-Enhancing Technologies and Responsible AI for tabular ML to building hardened, modular libraries that scale Multi-Agent Orchestration for millions of users.I thrive at the intersection of complex systems and cutting-edge research, building the infrastructure that makes foundation models useful at scale.
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.