Software Engineer, Infrastructure

📋 Job Overview

Build the software backbone letting AI reason about and experiment on the physical world. Dunia is building AI for one of the hardest unsolved problems in science: turning materials discovery from an academic, trial-and-error process into a programmable, scalable discipline. At Dunia software doesn’t just serve users, it coordinates robots, experiments, simulations, and learning systems that interact with matter itself. As Software Engineer, Infrastructure, you will own the systems that make this possible. Your work sits directly between scientific intent and physical execution: if infrastructure is unclear, brittle, or slow, discovery stalls. If it’s well-designed, the entire system accelerates. This role is for an engineer who wants their software decisions to have real, measurable consequences in the world.

🏢 About Dunia

Dunia is building AI for one of the hardest unsolved problems in science: turning materials discovery from an academic, trial-and-error process into a programmable, scalable discipline. At Dunia software doesn’t just serve users, it coordinates robots, experiments, simulations, and learning systems that interact with matter itself.

🎯 The Role

As Software Engineer, Infrastructure, you will own the systems that make this possible. Your work sits directly between scientific intent and physical execution: if infrastructure is unclear, brittle, or slow, discovery stalls. If it’s well-designed, the entire system accelerates. This role is for an engineer who wants their software decisions to have real, measurable consequences in the world.

✅ Key Responsibilities

• Production systems that matter
• Backend services in Python (FastAPI, Pydantic) that orchestrate real scientific workflows
• Data pipelines that turn raw experimental output into usable signal
• Tools people rely on daily
• Internal dashboards and interfaces (React, TypeScript, Tailwind, Streamlit) used by scientists and engineers
• Systems where usability and correctness directly affect productivity
• Infrastructure as a first-class product
• Containerized environments and CI/CD pipelines that keep the system moving
• Observability that tells you what’s happening before users complain
• Reliability improvements driven by actual failure modes, not hypotheticals
• The full loop: software → compute → science
• Job orchestration and ML system support
• Integration with SLURM-based simulation clusters
• Infrastructure decisions informed by real workloads, not abstract benchmarks