Software Engineer, AI

Magical

full-time

Posted on: 9/2/2025

Location: 🇺🇸 United States

✨ AI Apply

💰 $75,000 - $120,000 per year

Entry Level

AirflowJavaScriptNext.jsNode.jsPostgresPythonTypeScript

About the role

At Magical, we empower organizations to automate complex, manual workflows; building a brand new product with a founding engineer opportunity.
In a few months the product hit product-market fit, closed six-figure deals, and the existing product has generated millions in revenue.
Work in a small, nimble team tackling challenging, ambiguous problems at the cutting edge of technology.
Process Mapping & Discovery: interview stakeholders and translate business processes into step-by-step logic that can be automated.
Agent Building: prompt, code, and configure AI workflows; iterate on prompts, logic, and integrations to achieve production-ready agents.
Testing & Debugging: run test cases, identify failure points, debug agent behavior, and optimize workflows for speed, accuracy, and reliability.
Rapid Experimentation: deliver MVPs within hours or days; balance "done fast" with "done well" and iterate based on feedback.
Continuous Curiosity: ask questions to uncover bottlenecks and propose additional workflows or automations.

Education: BS in Computer Science (or near completion, recent grad).
Proficient in at least one scripting language (Python, JavaScript preferred).
Comfortable with APIs, data structures, and writing clean, efficient code.
Strong ability to translate ambiguous business processes into logical workflows.
Exceptionally fast learner who thrives in ambiguity.
Gets shit done: bias to execution, speed, and iteration.
Inquisitive and resourceful; digs deeper when something doesn’t make sense.
Thrives in a startup environment with high expectations and rapid feedback cycles.
Nice-to-have: Experience with automation tools (Zapier, n8n, Airflow, LangChain, etc.).
Nice-to-have: Prior internship or project experience building agents, chatbots, or workflow automation.
Nice-to-have: Knowledge of LLM prompting techniques.
Tech stack: TypeScript; Backend: Node, AI SDK, Hono, Drizzle, Postgres (neon), Railway; Frontend: Next.js, Shadcn, Tanstack, Tailwind, Zustand, Vite.