Magical is an agentic automation platform bringing state-of-the-art AI to healthcare—delivering AI agents that actually work in production.

We're building "AI employees" that automate the repetitive, time-consuming workflows that slow teams down. Our focus is healthcare—a $4 trillion industry buried in administrative complexity—where we automate claims processing, prior authorizations, and eligibility checks, enabling providers to focus on what matters: patient care.

Our Traction

The shift to agentic automation in healthcare is inevitable, and we're leading it:

- Scaled from $1M to $4M ARR in the first 6 months of our new agentic platform

- Accelerating growth with customers expanding into new workflows before renewal

- 7-day proof-of-concepts that demonstrate real value fast

- Self-healing automations with production-grade reliability

Unlike many AI companies making bold promises, we ship reliable solutions that deliver measurable results. We're backed by Greylock, Coatue, and Lightspeed with $41M raised. Our founder, Harpaul Sambhi, is a second-time founder who successfully sold his first company to LinkedIn.

What You'll Do

As an engineer at Magical, you'll:

- Shape the foundation of our agentic automation platform

- Build the future of healthcare automation alongside a high-agency team of 15 engineers

- Ship features that directly impact how healthcare providers serve patients

- Work at the cutting edge of what's possible with agentic AI, building and evolving how multiple agents collaborate to solve complex workflows

- Build rigorous verification systems, evaluations, and development experiences that guarantee our agents are doing the right thing every time

About the Role

As a Backend Engineer on the Agent Platform team, you'll build the foundational systems that power our AI agents. You'll design and implement core infrastructure—from agent orchestration and state management to model integration and evaluation pipelines—that enables reliable, scalable AI-powered experiences.

This role is for engineers who love building platforms and systems that others build on top of. You'll have complete ownership over critical backend infrastructure, working deep in the stack to solve complex distributed systems challenges. Engineers here are trusted to make architectural decisions, move fast, and deliver foundational capabilities that unlock new product possibilities.

In this role, you will

• Design and build core platform systems for agent orchestration, memory, and state management

• Develop robust infrastructure for integrating and evaluating the latest AI models at scale

• Build APIs and abstractions that enable product teams to quickly ship new agent capabilities

• Own critical backend services end-to-end, from architecture through deployment and monitoring

• Identify platform opportunities that unlock new agent behaviors and product features

• Work directly with the founding team to shape the architecture of our agentic platform

Your background looks something like this

• Have 4+ years of backend engineering experience, with strong proficiency in TypeScript/Node

• Deep understanding of distributed systems, databases, and asynchronous architectures

• Strong bias towards action and embody "show > tell"—you ship systems and iterate quickly

• High degree of agency: you effectively prioritize, unblock yourself, and drive projects forward without much outside input

• Value taking ownership and responsibility for your systems and infrastructure

• Studied Computer Science (or related field), or dropped out to build stuff

• Located in SF or willing to relocate

Even better

• Prior experience building platform or infrastructure systems at scale

• Experience integrating LLMs or AI models into production backend systems

• Background in real-time systems, event-driven architectures, or workflow engines

• Track record of building foundational systems that enabled teams to move faster

We're building the best self-serve agentic automation platform for the healthcare industry—and we're just getting started. Come join us.