Controversial employer on your resume? How to list it

Why are surveillance tech engineers suddenly job-hunting?

Because the backlash against the product reached the payroll. In early October 2026, reporting confirmed that Flock Safety — maker of the AI-powered license plate readers used by thousands of US police agencies — would shed roughly 18% of its approximately 1,500-person workforce, about 270 roles, after opening a voluntary separation program in September. The company described it as positioning itself for the next phase of growth and declined to say whether its chief executive would take a pay cut. The commercial backdrop is harder to spin: more than fifty towns paused or cancelled deals this year, Florida banned plate readers on state highways, several police departments let contracts lapse, and lawmakers pushed to require warrants before anyone queries plate data.

If you're one of those 270 people, you now have a very specific CV problem. Your last two or three years of work are genuinely strong — real-time object detection, OCR under terrible lighting, quantized models running on Jetson-class hardware, multi-camera tracking, fleet management for tens of thousands of deployed devices. But your CV probably describes all of it in the company's language: public safety outcomes, crimes solved, agencies onboarded. That framing was useful internally and it's a liability externally, because it asks every reader to have an opinion about surveillance before they get to have an opinion about your engineering.

Here's the reframe that matters. Your problem isn't moral, it's retrieval. A recruiter spends somewhere between seven and thirty seconds on a first pass, and an applicant tracking system reads your file as plain text before any human sees it. Neither one is weighing the ethics of automated plate recognition. They're looking for the terms in the job ad. Autonomous vehicles, retail analytics, industrial inspection, medical imaging, warehouse robotics, agritech and sports analytics all need exactly the skills you built — low-latency inference, edge deployment, data pipelines, model evaluation at scale. The work transfers cleanly. Only the product story doesn't.

Should you keep a controversial employer on your CV?

Yes. Keep the legal company name, your real job title and accurate start and end dates. Deleting a two-year role or hiding it behind "a technology company in public safety" buys you nothing and costs you plenty: recruiters routinely cross-check CVs against LinkedIn, employment verification is standard in background screening, and an unexplained gap draws far more suspicion than a company somebody dislikes. "Company confidential" is defensible only for a current employer during a genuinely discreet search, and even then it reads as evasive to most hiring managers. The moment a verification call contradicts your document, you stop being a candidate with an awkward employer and become a candidate who misrepresented their history.

What you do control is proportion. Right now your CV probably gives that employer the most vertical space, the mission statement and a bullet list written in product-marketing voice. Shrink the company to a single neutral descriptor — something like "venture-backed computer vision company, ~1,500 staff, large-scale camera networks and edge inference" — and give the space back to your engineering. Drop the taglines. Drop any sentence that argues the product was good or necessary; your CV is not the place to litigate that, and a reader who disagrees will simply stop reading. One factual line of context is enough for anyone to understand the environment you worked in.

Now the contrarian part: most hiring managers care far less than you think. Engineers worry about this for weeks and then discover that the interviewer's only question was whether the latency numbers were real. The people who will genuinely hold it against you are a small, identifiable group — privacy advocacy organizations, some civil-liberties-adjacent nonprofits, a handful of European teams with strong data-protection cultures. That's a few percent of your market. Don't rebuild your entire CV around the minority who'd reject you; write for the majority who need somebody who has shipped vision models to tens of thousands of devices in the field and kept them running.

Hebrew note aside, the same logic applies in any market: competence is portable, product politics aren't.

How do you rewrite bullets around the tech, not the product?

Start every bullet with the engineering action and end it with a measured result, then mention the product only as deployment context — never as the point. "Helped police agencies solve more cases with AI camera coverage" tells a hiring manager nothing portable and invites a political reaction. "Cut p95 inference latency 38% on resource-constrained edge devices by quantizing a YOLO-based detector, sustaining real-time throughput across 40,000 deployed units" tells them you can do the job in front of them. Same work, same honesty, completely different reading. Swap in your own verified numbers — model accuracy, mAP, frames per second, device fleet size, uptime, cloud egress saved, annotation throughput, build or retraining cadence.

Then rebuild your skills block so a parser can actually index it. Group the terms the way practitioners name them: languages (Python, C++), frameworks (PyTorch, TensorFlow, OpenCV), model families (CNNs, vision transformers, YOLO, Mask R-CNN), deployment (CUDA, TensorRT, ONNX, Jetson, Docker, Kubernetes), and data tooling. Use the exact spellings from the job ad, because a system matching on strings doesn't know your abbreviation and the synonym are the same thing. A single-column layout, standard headings and no text boxes or icons keep all of it readable; two-column templates quietly scramble the order in extraction. The software engineer CV guide walks through the structure in detail.

One more discipline: cut anything that only made sense inside that company. Internal tool names nobody else uses, team acronyms, the customer logos, the headcount of the agency onboarding team, awards tied to the product's mission. Replace each with the generic capability underneath — "internal labeling platform" instead of the codename, "regulated public-sector customers" instead of named departments. If you signed an NDA, describe the mechanism and the scale without the specifics; that's standard practice in defence and security engineering and no reasonable interviewer will push past it. When you've finished, run the file through a free CV analysis and see whether the ATS compatibility and clarity scores back up what you intended.

Aim for one page if you're under eight years in, two at most beyond that.

How do you answer "why did you work there?" in an interview?

Answer in three sentences: what drew you to the technical problem, what you learned, and why you're moving on. Something like: "I joined because real-time vision on constrained hardware is genuinely hard and they were doing it at a scale almost nobody else was. I learned a lot about reliability when your model runs on forty thousand devices you can't physically reach. The company's direction and mine diverged, and I'm looking for somewhere I can apply the same engineering to a product I'm more excited about." Then stop talking. Don't defend the product, don't trash it, don't offer a position on surveillance policy you weren't asked for.

Expect a sharper version of the question too, especially from mission-driven teams: "How did you feel about what the technology was used for?" The honest, non-defensive answer works best — acknowledge that the debate is real, say what you did or didn't have influence over, and describe what you'd want to be different. Interviewers are testing judgement and self-awareness, not looking for a confession. What sinks candidates is either extreme: a loyal recital of the company's talking points reads as a values mismatch, while bitterly dragging a former employer reads as someone who'll do the same to them in two years.

Resist one specific temptation: volunteering the layoff as an apology. "The company let 18% of us go after the backlash" is a fact you can state plainly if asked about your availability or your reason for leaving, and a group reduction is a clean, verifiable explanation that carries no personal stigma. But leading with it frames the whole conversation around a company's troubles rather than your work. Say it once, in neutral language — "my role was cut in a company-wide reduction in October" — and steer back to the systems you built. Rehearse all of this out loud twice before your first screen; it comes out stiff the first time.

Where should you apply, and how many CV versions do you need?

Target eight to twelve well-matched roles a week with a tailored CV rather than blasting fifty generic applications, and build three or four base variants instead of one. Engineers leaving a surveillance company tend to fan out into four natural lanes: autonomous systems and robotics, retail and physical analytics, industrial and infrastructure inspection, and medical imaging. Each lane values a different slice of your experience. The autonomy lane wants sensor fusion, tracking and latency. Medical imaging wants segmentation, data quality and regulatory rigour. Retail wants throughput, cost per camera and privacy-preserving design — which, ironically, is the part of your résumé that story makes most credible, not least.

Tailoring beats volume because the first filter is literal. Paste the job description next to your CV and check which of its named technologies, model families and deployment targets actually appear in your file; the missing ones are your edit list. HRLens's job-targeted analysis does this directly — paste the job text (copy it rather than linking, since LinkedIn and Indeed block automated fetching), and it returns the skills you're missing against that ad plus ATS-friendly rewritten versions of your CV in six templates that mirror the ad's own vocabulary. If your current document is a patched-up file from three jobs ago, it's faster to rebuild it with an AI CV builder than to keep editing around the damage.

Then work your network harder than your applications. When 270 people leave one company in a month, you have an instant alumni cohort — former colleagues land somewhere, and a referral from one of them skips the whole screening problem you've been worrying about. Message every ex-teammate who's already moved, be specific about what you want, and ask for an introduction rather than a job. Tell two or three recruiters who specialize in computer vision or embedded ML exactly which lane you're targeting; they'll pre-frame your background before a hiring manager ever reads it. Track every application in a simple spreadsheet with the date, the variant you sent and the follow-up date. Six weeks is a normal search. Nine isn't a verdict on you.

Frequently asked questions