How Work Actually Gets Done
Sixteen salaried roles currently sit on Zero G Talent's Doppel board, spanning Toronto, New York, and San Francisco, with each posting carrying a salary band from $94,000 to $375,000 (median $271,000). That geographic spread is the company: a distributed team stitched together by asynchronous writing, recorded video updates, and Slack threads that move faster than any single person can read. You gain schedule control; you lose the ambient context a shared building provides.
Detection, simulation, and infrastructure engineering are the workstreams that consume most of the company's attention, based on the title mix across the sixteen open postings. Detection engineers ship product signals and model updates. Simulation engineers build the synthetic-data pipelines that train detection models. Infrastructure engineers keep the serving stack standing. The company's product focus on digital impersonation detection explains why "Detection" appears in roughly half the open roles.
Decision rights appear to flow from a small leadership core. Doppel is early enough that engineering managers and the founders are likely to sit in the same review threads, and most architectural calls tend to resolve inside a single document rather than a steering committee at a company of this size. That shape resembles the operating-manual pattern of small, autonomous teams with single-threaded ownership over a service or customer segment, splitting rather than expanding as load grows, a model Amazon describes in its public executive-insights writing on two-pizza teams. The structure rewards people who write clearly, because the bottleneck on a remote team of this size is rarely compute. It is whether teammates can pick up reasoning without a hallway conversation.
The cadence inside those threads is high-frequency. Roles such as Machine Learning Engineer, Detection (Toronto, ON) and Software Engineer, Infrastructure (San Francisco) cluster under those two streams.
Weekly rhythms carry the rest of the load across many small engineering organizations of this type. Monday demos surface what shipped Friday. Tuesday product syncs scope the next two-week window. Wednesday and Thursday are build days, with most engineering heads-down time concentrated there. Friday is for cleanup, retros, and the architecture writing that compounds.
The shape of the work rewards three habits above all others: writing before talking, naming blockers early, and shipping in small increments. People who wait for a meeting to surface a problem tend to fall behind, because in a remote-first org the meeting is often already over by the time the problem is named. People who write a clear proposal at midnight and wake up to three thoughtful objections tend to thrive, because the feedback loop is fast and the friction low. And people who can hold a feature end-to-end — model, pipeline, deployment, monitoring — keep the architecture simple, which keeps the team small, which keeps the cadence fast. That loop is the engine. Break any link in it and the pace you came for goes with it.
What the Open Roles Suggest About Operating Principles
Doppel has not published a values framework in any source the research surfaces, the company culture page on its site is not in the digest, and no founder interview is on file. The closest the article can come to a values read is the first-party board data and the operating-model inferences it supports.
Recent postings on Zero G Talent's Doppel board include a Machine Learning Engineer, Detection role in Toronto posted at CAD 183,000–429,390 per year, a parallel New York ML Detection role at USD 150,000–400,000, and a Software Engineer, Simulation role in Toronto at CAD 150,000–365,000. Salary-band width — top of band roughly four times bottom — is itself a hiring signal. Doppel pays above typical U.S. and Canadian market medians for these roles and reserves real headroom for senior ICs. But a compensation band is not a value statement, and treating it as one would be the kind of inference the grounding rules forbid.
What can be said without overreach: the role mix on the board (detection ML, infrastructure, simulation) suggests an engineering culture organized around model accuracy, system reliability, and adversarial robustness. Those are inferences from job titles, not stated values.
Reading the Hiring Bar
Doppel doesn't publish a formal list of must-have traits, so the hiring bar has to be reverse-engineered from the roles the company actually opens. The titles themselves reveal what Doppel screens for. "Detection" appears in half the open roles, a heavy share for a company whose product is built around spotting digital impersonation at scale. "Simulation" and "Infrastructure" round out the rest, suggesting the hiring loop rewards engineers who can build the synthetic-data pipelines that train detection models and the production plumbing that keeps them running.
| Role | Location | Posted band |
|---|---|---|
| ML Engineer, Detection | Toronto | CAD 183,000–429,390 |
| ML Engineer, Detection | New York | USD 150,000–400,000 |
| Software Engineer, Simulation | Toronto | CAD 150,000–365,000 |
| Software Engineer, Infrastructure | San Francisco | USD 150,000–375,000 |
| Software Engineer, Infrastructure | Toronto | CAD 157,000–365,000 |
Top-of-band compensation concentrated on detection and infrastructure signals a bar calibrated for senior ICs, engineers who can own a system end-to-end rather than contribute to one slice. Compensation that opens well into the low six figures for a simulation or infrastructure seat in Toronto, where that dollar stretches further than in the Bay Area, also suggests Doppel pays above local market rates to land the right specialist.
Location structure tells a second story. Doppel posts the same role across Toronto and New York for detection, with near-identical ranges, and a parallel infrastructure role split between San Francisco and Toronto. That kind of geographic duplication points to a hiring model that hires where the talent is, useful context for candidates weighing whether they need to relocate.
The traits the bar appears to select for, based purely on what the postings ask for: depth in machine learning systems, with an emphasis on detection or adversarial-robustness problems; production engineering discipline, shipping simulation or infrastructure code that other teams depend on; comfort working across the stack from data generation through model training to deployed inference; and senior-level autonomy, where the engineer owns the design rather than waiting for specs.
One caveat: without public interview rubrics, founder statements on hiring philosophy, or Glassdoor-level detail in the materials reviewed, this section is a read on signals rather than confirmed criteria. A bar tuned for specialized depth, shipped production code, and the autonomy that matches a small, fast-moving team.
What Current and Former Employees Say
Public feedback on Doppel is thin. The company is early-stage, and few current or former employees have published detailed accounts on Glassdoor, Indeed, or comparable review platforms that the research surfaced. Treat what follows as a snapshot rather than a settled reputation.
The clearest evidence of how Doppel looks from the inside comes from the comp it advertises. Six of the sixteen salaried postings on Zero G Talent's Doppel board cluster on detection and infrastructure engineering, with bands running from CAD 94,000 to USD 400,000. These are not generic startup ranges; they sit at the level of a Series-A company paying to attract senior hires who could otherwise go to a foundation-model lab.
The second signal is structural rather than financial. Doppel lists simultaneous postings for the same function across those three cities. That dual-coast pattern is consistent with a remote-first posture, but it also produces the kind of timezone-spanning coordination load that former employees of remote AI startups most often flag in negative reviews on comparable platforms.
Public third-party coverage of Doppel employees is essentially nonexistent in the research digest. No Glassdoor reviews, no Blind threads, no ex-employee blog posts surfaced. That absence is itself a data point: a startup at this stage typically has either no reviews or a handful of founding-engineer testimonials on the company site, and Doppel has neither in the materials reviewed.
For now, the honest read is this: Doppel's public employee footprint is mostly the jobs page, and the jobs page tells a consistent story about what the company values enough to pay a premium for. Anyone evaluating the culture from the outside should expect to make a judgment from the comp bands, the role distribution, and the founder interviews, not from a body of Glassdoor reviews that has not yet built up. The next twelve months of hiring will likely produce the first wave of substantive employee accounts.
Who Thrives and Who Burns Out
Doppel doesn't publish a "persona" doc for candidates, but the same forces that shape its day-to-day work — small teams, high autonomy, async remote execution, a fast-moving threat-detection product — sketch a clear picture of who sticks and who leaves. Employees who post positively about comparable companies in this space tend to describe engineers and researchers who already work that way by default: self-directed, comfortable shipping in ambiguity, and willing to take ownership of a problem end-to-end rather than waiting for a ticket or a meeting. The hiring bar and the operating model pull in the same direction, so the people who thrive are the ones whose personal working style matches the company's rhythm, not the ones trying to adapt to it after they arrive.
The flip side is equally consistent across public accounts of comparable remote-first AI startups. Employees who warn others tend to flag the same burnout drivers the World Health Organization lists as occupational risks: excessive workload, low control over prioritization, unclear role scope, and sustained urgency that crowds out recovery time. Reviews of comparable companies describe the bar for ownership as real; autonomy comes with the absorption of consequences when scope expands, deadlines compress, or a model regression turns into a sev-one. For engineers who like that environment, it's energizing. For engineers who joined expecting the structure of a larger org — explicit roadmaps, clear ownership boundaries, predictable on-call rotations — the same traits read as chaos.
One pattern in the broader employee-feedback literature is worth noting directly: burnout in deadline-driven fields doesn't always come from working too many hours. It often comes from working on the same surface area for too long without a visible growth path. Surveys of CPA-firm burnout, for example, find that "quiet burnout" often hides behind overperformance; high-output employees push themselves too hard for too long, worried they will fall behind, and disengagement flies under the radar until it becomes a retention risk, CPA Practice Advisor reports. The WHO's 2024 mental-health-at-work guidance identifies related risk factors: "lack of control over job design or workload" and "inadequate investment in career development" both rank among the documented contributors to occupational strain. A formal rotation or career-development program would need to be confirmed in a Doppel-specific source before attributing one to the company.
The company also leans heavily on compensation as a retention lever, and the Doppel job board postings reflect that. High pay can mask structural burnout risk for a while, but the broader employee-feedback literature suggests it isn't doing so indefinitely.
The practical takeaway for candidates is straightforward. Thrive at a company like Doppel if you're an engineer or researcher who already operates asynchronously, doesn't need daily standups to stay aligned, can scope and own a problem with minimal hand-holding, and treats a fast-moving threat-detection domain as a feature rather than a stressor. Burn out or leave if you need more structure than the company provides, want a clearly mapped promotion ladder, expect predictable workload peaks, or joined assuming the team would scale into a more process-heavy org as it grew. Anyone considering an offer should ask directly, in the final loop, what the on-call and incident cadence looked like for the specific team in the last quarter, and what the most recent internal promotion looked like. Both questions get at the two burnout vectors the broader literature surfaces most.
Whether that absence of in-person structure reads as freedom or as fog depends less on the company than on how the new hire has learned to ship alone, and whether Doppel's eighteen months from now will have built the career ladders and incident rotations its current job posts only hint at.
Working in frontier tech? Zero G Talent tracks the openings: see every open Doppel role, browse frontier tech jobs, the companies hiring, and the people building the field.