Markel AI Automation Intern

Markel•Woodland Hills, CA

About The Position

Markel is seeking an AI Automation Intern to help build practical tools that reduce repetitive work for underwriters, claims handlers, and operations teams. Projects include AI-assisted underwriting workbenches, automated email flows, and integrations between existing systems. Work typically moves from discussion with a business leader to a proof of concept in days, followed by testing, hardening, and handoff. The intern will be the second person on this work, with opportunities to learn specialty insurance and apply AI safely in a regulated business.

Requirements

Bachelor’s degree or anticipated graduation by December 2026.
Working skill in at least one programming language, ideally Python or JavaScript, demonstrated through coursework, an internship, or personal projects.
Basic understanding of web APIs, including HTTP requests, JSON, and authentication.
Strong Excel skills, including lookups, pivot tables, and formulas, and comfort working with large data sets.
Clear writing and ability to explain work and its purpose to people who do not code.
Curiosity about AI and the habit of checking its answers rather than trusting them.
Good judgment with data, including asking before sharing or sending customer and broker information.
U.S. work authorization; Markel does not provide visa sponsorship now or in the future.

Responsibilities

Build and improve proof-of-concept tools, including browser-based workbenches for reading submissions, pricing risks, and producing quotes.
Connect systems through APIs by sending requests, handling authentication and tokens, mapping returned data, and managing errors.
Write Python scripts to turn spreadsheets, PDFs, and Word documents into clean data and to generate finished documents from data.
Build and maintain Power Automate flows, SharePoint lists, and Copilot Studio agents for email triage, document review, and reminders.
Use large language models by writing prompts and extraction schemas and creating test sets to evaluate accuracy and cost.
Test tools before they are relied upon, using automated browser checks, unit tests, and comparisons against source data.
Deploy tools on the Databricks cloud data platform while managing access, secrets, and costs.
Write setup notes, runbooks, user guides, and change logs to enable others to take over projects.
Meet with underwriters and business leaders, capture their needs, and turn them into clear requirements.