Magical

Staff Software Engineer

San Francisco, California, United StatesFull timeStaffPosted 12 days ago
Apply on Magical →

Sign into see who you know at Magical.

ABOUT MAGICAL

Magical is an agentic automation platform revolutionizing healthcare with state-of-the-art AI—delivering production-ready AI agents that drive real results at scale.

 

We believe the future of healthcare lies not in replacing people, but in multiplying their best qualities. Magical empowers healthcare organizations to scale their top experts across every workflow, ensuring more accurate, efficient, and compassionate operations.

 

We're building "AI employees" to automate repetitive, time-consuming tasks while keeping humans in control. Our focus is healthcare, a $4 trillion industry mired in administrative complexity. By automating processes like prior authorizations and medical coding, we enable providers to concentrate on patient care.

 

OUR MOMENTUM

- Customers rapidly expand from initial workflows to broader use cases after seeing reliable production results.

- 7-day proof-of-concepts deliver real value quickly—an industry where months are typical.

- Self-healing automations maintain production-grade reliability at scale.

 

Magical is backed by Greylock, Coatue, and Lightspeed, with $41M raised. Our founder, Harpaul Sambhi, previously sold his first company to LinkedIn.

 
 

ABOUT THE ROLE

As a Staff Software Engineer, you’ll shape the technical direction of Magical’s AI agent systems. You’ll drive product delivery, elevate engineering standards, mentor peers, and stay hands-on with the code.

 

Magical orchestrates complex, high-stakes healthcare workflows involving queues, run state, browser automation, model calls, system evaluations, observability, and human-in-the-loop recovery.

 

You'll transform messy customer inputs—SOPs, decision trees, portal behaviors, edge cases—into robust automations. This means building systems that draft workflows, test them, catch failures, suggest improvements, and trigger human intervention when necessary. You'll also focus on runtime reliability, workflow pathways, evaluations, and speci...