Forward Deployed Physical AI Engineer
Location: United States - field-based | Travel: ~80% Full-time | Overview.ai
The Opportunity
Overview.ai builds physical AI systems, computer vision and machine learning deployed directly into production lines, wired into the same PLCs and equipment that run the factory. We're not a software vendor bolting a dashboard onto someone else's line. Our systems see, decide, and act inside real manufacturing processes, running today in production at companies like Tesla, SpaceX, and Amphenol.
We're hiring a Forward Deployed Physical AI Engineer to be the person who takes a system from "this could work here" to "this is running production, unattended, at spec" - inside some of the most demanding manufacturing environments in the country. This is not a support role. You will design the imaging, AI, and integration approach for each deployment, own the technical relationship with the customer's engineering team, and carry the outcome from first site visit through steady-state production.
Why This Role Matters
Physical AI only works if it survives contact with a real factory: vibration, inconsistent lighting, legacy PLCs, production schedules that never stop for you. That's the hard part, and it's the part most AI companies get wrong because they never put engineers on the floor. You will be that engineer. What you learn on-site, where the models break, where the optics fail, where the PLC integration gets fragile, feeds directly back into what we build next. You're not just deploying the product; you're one of the primary sources of ground truth for what physical AI needs to become.
What You'll Own
Solution design. Diagnose the customer's actual manufacturing problem, then design the camera, lighting, and AI configuration to solve it, not just install a preset.
Systems integration. Integrate Overview.ai systems with customer PLCs (Allen-Bradley, Siemens, and others), mapping I/O, configuring communication protocols (EtherNet/IP, PROFINET), and validating connectivity without disrupting existing controls.
Production validation. Take systems from bench-tested to production-proven under real constraints: line speed, part variation, vibration, lighting drift.
The customer relationship. Be the primary technical contact for plant engineers, controls teams, and often plant leadership, from first discovery call through live production and beyond.
Product feedback. Bring field reality back to our product and engineering teams — what breaks, what's missing, what the next version needs to handle.
What Makes This Role Different
This role gets confused with field service or applications engineering. It isn't either.
Field service follows a checklist. This role designs the solution. Applications engineering typically supports a pre-sold, pre-specified product. Here, you're often the first engineer figuring out what the right solution even is for a given line, and your judgment calls become the deployment. You'll work with production hardware, real electrical and controls problems, and AI systems simultaneously, in environments where "it works in the lab" means nothing until it works on the line.
You'll also see more manufacturing environments, PLC architectures, and production challenges in a year than most controls or automation engineers see in five: because you're not embedded in one plant, you're moving between many.
What We're Looking For
Strong candidates tend to show up in one or more of these forms:
You've integrated hardware or systems into something that already existed and worked a production line, a robotics platform, a competition vehicle, without breaking what was there.
You've gone deeper than your job required: debugged something nobody assigned you, built a side project involving sensors/controls/automation, or led a robotics, FSAE, or similar technical team.
You're comfortable being the only person in the room who understands both the AI/vision side and the electrical/controls side of a problem.
You can explain a technical failure to a plant electrician and to a VP of Operations in the same afternoon, in language each of them trusts.
You've operated independently in ambiguous, high-stakes situations: a live production line, a competition deadline, a research deployment where there was no playbook.
Must-haves:
Hands-on experience integrating hardware/software systems with PLCs or industrial controls (professional, research, or high-caliber project/competition experience)
Working understanding of industrial communication (EtherNet/IP, PROFINET, or equivalent) and device description files (EDS/GSD/GSDML)
Comfort with electrical/controls fundamentals: reading schematics, using a multimeter, diagnosing signal and connectivity issues
Track record of ownership: solo or lead responsibility for taking something from broken/undefined to working
Genuine willingness to travel ~80% and work independently on customer sites for multi-day to multi-week stretches
Preferred:
Exposure to Allen-Bradley or Siemens platforms specifically
Computer vision, imaging, or optics experience (cameras, lighting, lensing)
Prior customer-facing technical role
Technical Foundations
You don't need to be a PLC expert on day one, but you should already be fluent in the fundamentals: how PLCs talk to devices on a network, what EtherNet/IP and PROFINET are for, why device description files (EDS/GSD/GSDML) matter, and what it actually looks like to add a new device to an existing production network without taking the line down. If none of that is familiar territory, this role will be a steep and probably frustrating climb, better to find that out now than three weeks into a deployment.
Travel Expectations
Direct version: this role is ~80% travel. That's the job, not an occasional feature of it. You'll spend most weeks on customer manufacturing sites, working around production schedules that don't pause for you.
We're not going to pretend that's easy or glamorous. It's demanding, and it's not the right fit for every strong engineer.
What it does get you, that a remote or single-site role can't: you'll walk the floor of more advanced manufacturing operations in a year than most engineers see in a decade. You'll own full deployments end to end instead of advising from a distance. You'll build the kind of judgment and customer-facing credibility that normally takes years of tenure to earn. And you'll be one of the people this company relies on to understand what physical AI actually needs to work in the real world.
If travel is a dealbreaker, this isn't the right role and that's a legitimate answer, not a failure.
Career Growth
This role is a fast, high-density way to build expertise across AI, computer vision, controls, and manufacturing systems simultaneously; exposure that's hard to get anywhere else, because most roles specialize in one of these, not all four. Engineers who excel here are positioned for technical leadership, regional/deployment leadership, or product roles where field-earned judgment is exactly what's missing from a typical product team. As Overview.ai scales, we expect this role to be a primary feeder into senior technical and leadership positions, including roles with materially reduced travel over time.
About Overview.ai
We build physical AI systems for manufacturing quality control, designed to run inside real production lines, integrated with existing PLCs and equipment, not bolted on as a separate tool. We're deployed in live production today with companies including Tesla, SpaceX, and Amphenol. We're a small, fast-moving team: decisions happen quickly, ownership is real, and the work you do is directly visible in whether a customer's line runs correctly.