Software Engineer

Job description

**Radical AI** is replacing an R&D process that currently takes 10+ years and $100 million to produce a single discovery. Our self-driving lab platform combines AI with autonomous robotics to run experiments, analyze results, and iterate—continuously, without human bottlenecks. For industries like aerospace, automotive, defense, energy, manufacturing, semiconductors, and space, that means breakthroughs in weeks instead of years. ### The Role This is a generalist software engineering role on the team building the core platform that powers our autonomous lab. Depending on where you plug in, you might own backend services, agent tooling, data and orchestration pipelines, internal platforms, or the infrastructure that keeps our systems observable and reliable in production. What's non-negotiable: strong systems thinking, production instincts, and the ability to contribute across the stack. We're building software that controls and monitors physical systems in the real world — the bar for correctness and reliability is high. We work across a deeply cross-disciplinary domain (robotics, ML, experimental automation), and one near-term priority involves hybrid cloud/on-prem deployments in customer environments. Engineers who've navigated those constraints will hit the ground running. ### What You'll Work On - Lab backend: experiment definitions, sample path-planning, long running durable task execution - Data backend: database schemas, migrations, ETL pipelines, object storage + partition design - Internal platforms: developer tooling, SDKs, shared services, service templates - Observability and reliability: structured logs, metrics, tracing, production debugging (OpenTelemetry, Prometheus/Grafana) - Hybrid infrastructure: cloud + on-prem, containerization, orchestration, infrastructure-as-code - Agent capabilities and tooling: API integrations, code execution, scientific literature retrieval, workflow automation - Scientific workflow orchestration: Bayesian optimization loops, experiment scheduling, long-running job execution, retries, idempotency - Data pipelines for ingesting, transforming, and serving data to models and LLMs ### What We're Looking For - 5–6+ years of production software engineering experience; strong enough to design, build, and ship end-to-end - Fluency in Go and/or Python; additional stack experience (TypeScript, Rust) is welcome - Deep comfort with distributed systems: timeouts, retries, idempotency, partial failure, dead-letter queues, safe rollback - Experience with concurrent and asynchronous programming — event loops, cancellation semantics, bounded queues, task orchestration under failure - Solid networking fundamentals (especially important for potential on-prem deployment contexts) - Containerization and cloud deployment experience (Kubernetes, AWS); comfort debugging Linux systems - High ownership: you find problems before they find users and raise the bar for quality ### Nice to have - Experience with agentic systems or LLM workflows (tool-calling, context management, PydanticAI, LangChain) - Experience with enterprise self-hosted storage solutions (Ceph/Rook, Longhorn, WEKA, VAST, TrueNAS) - Ray framework experience - Familiarity with MongoDB and gRPC - CI/CD and DevSecOps experience (GitHub Actions, Gitlab Pipelines) - Familiarity with embedded protocols (serial, I²C, Modbus), device virtualization, or microcontroller firmware - Strong observability experience (Datadog, Prometheus, Grafana, ELK, distributed tracing) - Frontend experience (TypeScript/Svelte/React) — real-time interfaces, state management, structured APIs