Platform Engineer

Rimward is building the decision layer for physical security.

We start with remote outdoor sites such as utility-scale solar, where recurring intrusion, false alarms, and limited on-site presence make existing monitoring hard to operate well. Most serious sites already have cameras, procedures, and supervision teams. The missing piece is reliable signal.

Our goal is to turn existing camera networks into operator-ready alerts: fewer false positives, faster video verification, better evidence, and a tighter path from detection to response.

We care about field reality more than demo quality.

That means working on products that respect messy deployments: mixed camera quality, outdoor conditions, patchy connectivity, constrained infrastructure, and real operator workflows. We are opinionated, hands-on, and demanding about usefulness. A system that looks impressive but does not help someone make a better decision is not good enough.

We work in person in Paris while we sharpen product-market fit.

Rimward builds software that turns existing camera infrastructure into low-noise, operator-ready intrusion alerts for remote outdoor industrial sites.

The company is led by F-G Fernandez, a repeat founder with a strong computer vision background, who previously built and deployed vision systems in the field with Pyronear. We are now applying that same deployment-first mindset to physical security: building the software and systems layer that makes existing camera networks actually usable for fast, reliable intrusion detection on real sites.

We are looking for a platform engineer to help turn early pilots into a reliable product and internal platform. You will work on the backend and systems layer that make deployments repeatable: alerting pipelines, integrations, auth, observability, fleet management, admin tooling, and product infrastructure.

This is not a generic backend role. The product has to work across messy field realities: remote sites, uneven connectivity, mixed hardware environments, customer IT constraints, and workflows that bridge edge devices, video pipelines, cloud services, and human operators. A bug is not just a failed request. It can mean a broken deployment on a live site, poor operator trust, or a field issue that is expensive to diagnose.

Within your first 30 days, you should be able to understand the current deployment architecture, backend stack, and operational bottlenecks across pilots. Within 60 days, you should be shipping improvements that make deployments more repeatable and the system easier to operate internally. Within 90 days, you should own meaningful parts of the platform: reliability of alerting flows, internal tooling, deployment hygiene, and the productization of patterns discovered in the field.

You should enjoy building robust systems, reducing operational friction, and turning messy real-world constraints into clean product and infrastructure decisions. You should also be comfortable working with AI coding tools, while being strong enough to review, correct, and shape the output rather than just accept generated code.

Rimward sits at the intersection of computer vision, video systems, edge inference, and operator workflow design.

The core technical challenge is not just detecting events in video. It is producing signal that stays useful under real deployment constraints: outdoor scenes, variable lighting, weather, noisy backgrounds, uneven camera quality, and low-connectivity environments. Low false positives matter because the product only works if operators keep trusting the alerts.

Our stack is pragmatic and software-heavy: Python for backend and ML, modern video/data pipelines, and lightweight product surfaces for alert review and operations. We work across model performance, deployment, evidence generation, and infrastructure that can run both in the cloud and close to the site when needed.

1. Intro call (30 min)
2. Architecture interview (60 min)
3. AI-assisted build exercise (2–4 hours max)
4. Code review / debugging interview (60 min)
5. Founder final (45 min)