Staff Engineer

Evercam; United States; Remote; USD 120,000 - 190,000; Full time

Benefits offered by this job

PTO

1 extra PTO day per year served

National public holidays applicable in

Job summary

Evercam seeks a remote full-stack engineer to lead the development of AI-powered workflows for construction safety and efficiency.

You will design production-grade systems that integrate Elixir backends, Python AI services, Vue.js frontends, and PostgreSQL data stores, while owning the product journey end-to-end with high autonomy and revenue KPIs.

Qualifications

• Experience designing AI-powered systems at production scale.
• Ability to own the product journey from concept to production.
• Strong leadership and decision-making in autonomous teams.

Responsibilities

• Design and build agentic workflows to improve safety and efficiency on construction sites.
• Own production-grade systems and ensure reliability across distributed components.
• Collaborate with hardware, firmware, frontend, and backend teams to deliver end-to-end solutions.

Skills

Leadership

Product-minded

Full-stack engineering

Decision making

Tools

Elixir backend

Python AI services

Vue.js

PostgreSQL

Job description

Evercam unifies fixed cameras, drones, and 360° into a single platform to improve construction productivity, team collaboration, and visibility.

We are built for the most complex projects in the world, helping builders to run safer, more productive, and more profitable.

Job Description

This is a remote position.

About the Role

Evercam Reality Driven Intelligence powers modern construction on many of the world’s largest, most complex construction projects. With thousands of cameras deployed in over 30 countries, we are entering an exciting new period of turning these video streams into actionable data.

Safety is high on our customers’ priority list.

We’ve solved a lot of the necessary plumbing to make this into a very rewarding job for a talented, product-minded engineer to own the product journey from beginning to end.

We want you to design and build the agentic workflows that will increase compliance and reduce injuries and deaths in one of the most dangerous professions.

This is not a pure computer vision project, and it’s not a pure product role - it’s full-stack, high autonomy, high agency, engineering & product. Your KPIs will include revenue targets.

You will be supported by a team of experienced engineers across hardware, firmware, frontend, and backend. We have a modern, fit-for-purpose technology stack and can fill gaps where you have them.

We need a leader who can make decisions and build production-grade solutions.

About you

The right engineer must understand how to make models, agents, video streams, state, memory, and decision-making work together reliably at production scale.

About the app

Imagine an AI agent that can observe a live construction site, understand what's happening, decide what deserves attention, investigate across cameras and other information sources, and learn from feedback.

Not another AI chatbot. Not another object detection model. An intelligent system that operates in the physical world.

We're bringing together computer vision, multimodal models, agentic reasoning, real-time systems, and persistent memory to make this possible.

And we're looking for an exceptional engineer to lead the effort. Someone who wants to solve problems that don't have established playbooks, challenge how these systems are built, and turn ambitious ideas into reliable products.

If building intelligent systems that understand and interact with the real world sounds like the kind of challenge you've been looking for, we should talk.

Requirements

Our platform combines Elixir backend services, Python AI services, Vue.js and PostgreSQL, supported by established camera, video and media infrastructure.

You’ll build on this foundation. The engineering spans:

Computer vision and multimodal models: understand objects, activity, relationships and changes across footage.

State and memory: maintain current site understanding, event history, project knowledge and open investigations.

Agent-directed attention: decide what needs investigation, which evidence to gather and when to check again.

Tool-based investigation: examine surrounding footage, nearby cameras and project documents, with camera control where supported.

Reliable operation: durable execution, queues, retries, permissions, human review and evaluation of missed hazards, false findings, latency and cost.

You’ll help choose the models, agent framework, and memory architecture, and turn this product model into a reliable production system.

• PTO
  
• 1 extra PTO day per year served
  
• National public holidays applicable in your country