How work actually gets done, pace, structure, how decisions get made and who makes them
Tennr has tripled its revenue since its Series B in October 2024 and crossed into processing over 10 million patient records in 2025, per the company's own figures. That kind of compounding, in a healthcare category notorious for glacial sales cycles, tells you what the inside of the company has to look like: small teams shipping constantly, with no time for the committee work that kills most enterprise startups before they find product-market fit. The company's early-stage environment emphasizes rapid iteration and close-knit teams, which drives high autonomy but also creates intensity that some employees find unsustainable.
The flat structure shows up first in who actually works there. Tennr's New York office currently has 18 salaried roles open on the board, ranging from Product Engineer at $165,000–$190,000 to Enterprise Solutions Consultant at $205,000–$235,000. The titles are deliberately lateral; Senior Backend Software Engineer, Senior Product Engineer, Senior UX Engineer, Enterprise Account Partner all sit in roughly the same $190,000–$215,000 band. There is no "VP of X" layer visible in the public posting mix, and the median board salary lands at $175,000. That's the pay fingerprint of a company that hires individual contributors who own outcomes rather than managers who coordinate them.
That ownership extends to how decisions actually get made. Tennr ships proprietary large language models (it calls them RaeLM, a series of specialized models trained on medical documentation nuances), which means product, research, and engineering can't be separated by a roadmap meeting; they have to be in the same room, looking at the same edge cases. The company's LinkedIn feed has made the posture explicit: "Building our own models gives us the ability to fine-tune the reasoning of minute, edge-case tasks, so no patient is misrouted." When the model has to handle "checkboxes and messy handwriting" off a fax, the person who writes the prompt and the person who closes the customer ticket are often the same person.
Workflow runs against customer evidence rather than internal OKRs. One customer's experience absorbing 260 faxes in a single day without backlog growth: Amy Willis, Director of Complex Rehab at King Drug, told Tennr "We hit 260 faxes in one day… and were able to manage it." That sentence only makes sense if Tennr's platform is built to absorb 10x spikes without the customer needing to hire. The pace has a cost, and the company is honest about it. Tennr's own About-page language describes providers "drowning in piles of messy documents, ever-changing payer guidelines, and 45 minute phone tree wait times," and the company's job is to make that drowning visible and actionable in real time. The internal version of that posture is a workforce that has to keep up with a system processing millions of care pathways a month, launching two new product lines (Workspace and Eligibility & Benefits Management) within a single quarter of 2025, and onboarding senior operators like Maulik Shah (former health-system GM) and Bruna dos Santos (DME industry lead) to push into adjacent verticals without slowing core delivery. Fast iteration and tight feedback loops are how work gets done at Tennr; the same forces that triple revenue in two quarters are the ones that leave little room for the people running underneath them to coast.
What Tennr's stated values actually mean on a Tuesday
Tennr's stated values read like a founding-team manifesto, and the public record backs them up with unusually specific language rather than corporate filler. The company's own About page leads with three positions: meet providers where they are ("faxing isn't the problem"), practice AI transparency ("we built our own machine learning models and provide transparency into how they're trained, how they benchmark and what they do"), and pursue a problem the founders describe as "completely solvable": eliminating pre-visit patient delays. Co-founder Trey Holterman told Forbes the company was born from "one of the founders' ongoing struggle to get an appointment with a specialist to whom he had supposedly been referred," and the team's LinkedIn presence restates it flatly: "Tennr was founded with the purpose of eliminating patient care delays across the US."
The first operating principle that surfaces repeatedly in leadership commentary is build on top of what exists rather than replace it. Holterman said in a CareTalk: Healthcare. Unfiltered. Podcast interview that the team "weren't interested in replacing their faxes" and were "convinced that you could actually build their sort of dream solution with the faxes… people have had 20 plus years with solid database technologies to build around it and they haven't been able to." Tennr's platform "sits on top of existing EHR software, first reading and analyzing the incoming e-fax, and then entering the follow-up work," per Axios, and the company documented working "on top of" a pre-built intake tool at Noctrix Health instead of ripping it out. For employees, this principle translates into a working environment dominated by what the same CEO called "a very large integrations team" supporting "over 47 different EMRs." That's not a metaphor for cross-functional collaboration; it's a structural fact.
The second principle is vertical specificity over horizontal generality. Tennr's About page calls out a counter-move directly: "We're tired of software companies hand-waving the term 'AI' to mask their shortcomings." A LinkedIn post makes the same point more sharply: "Automation doesn't fix a broken workflow, it just makes broken faster." Tennr's leadership rejects the agent label in public commentary: "we don't say agents," Holterman said on the CareTalk podcast, describing in-house models built to do "the hard sort of reasoning steps." Forbes quoted an investor echoing the same framing: "Amidst the theoretically unbounded possibilities of AI, the Tennr team has impressed us with their unwavering focus on building applications solving specific, tangible problems for their customers."
The third principle is patient invisibility as a success metric. Holterman on the CareTalk podcast described the ideal outcome: "the experience of the patient, the patient almost never knows the tenor exists. The experience of the patient is simply by the time they're in the parking lot, they're getting a call and they're ready to go." Tennr's About page restates the goal in operational terms: "Solving the pre-visit patient processing problem will massively reduce patient delays and denials across the U.S. healthcare system." That invisible-patient framing sets an internal standard. What the team calls a 17% median throughput lift post-go-live, "over 50%" of denials "avoidable upfront," and the company's 2025 milestone of "over 10M patients" processed are all evaluated against whether the patient ends up seen faster, not whether the software is visible.
The fourth principle is ten-year obsolescence as a feature. Holterman added on the same show: "we can hopefully, you know, wipe our hands clean in maybe 10 years and say, 'Hey, we did it. Job well done, guys. We did a good job. Let's go home.'" That ambition is exactly what produces the early-stage intensity: a small team, post-Series A in March 2024, sitting on $25 million raised, processing 10 million documents a month and targeting "point C, point D, point E, and F" patient journeys while explicitly building a system they hope won't need to exist in a decade.
What the hiring bar SELECTS FOR, the traits and signals that get someone through
Tennr's hiring isn't built around credential theater. It's built around a specific kind of stubbornness: the willingness to keep digging into messy, ambiguous clinical paperwork long after a faster, less accurate answer would do. The company says so directly. Co-founder and CEO Trey Holterman told Healthcare Brew in February 2026 that Tennr is "laser focused" on the referral pipeline, deliberately avoiding the temptation to expand into scribing or capacity management. "I think that's a blessing to go really, really deep in an area, as opposed to all over," Holterman said. Candidates who generalize — who pitch themselves as comfortable across many healthcare AI problems — are signaling the wrong thing.
What the company rewards instead is depth over breadth, and a particular flavor of contrarianism about how AI gets applied to healthcare. Tennr's flagship model, RaeLM, is a vision-language system trained on 100 million anonymized healthcare documents, 2.3 billion distinct data fields and 8,000 sets of payer criteria, per a June 2025 Fierce Healthcare report. It was built because, as the company's October 2024 funding announcement put it, "nothing on the market, paid or open-source, was even close to hitting the accuracy requirements we needed" on tasks as granular as checkbox detection. Tennr's own LinkedIn copy calls the team "control freaks" about model behavior. The implicit hiring signal: don't come in arguing for a generic large language model wrapper; come in ready to defend why a custom, narrow system was necessary and what you'd build next.
That contrarian streak shows up in the founders' own backgrounds. Holterman, Diego Baugh and Tyler Johnson met as Stanford engineering students working on advanced AI and large language model research before founding Tennr in 2021 in New York. Baugh's motivation was personal; six-week GI appointment delays sent him to the emergency room in college. Holterman's came from watching his mother, a family-medicine practitioner, navigate what he called the "black hole" of referral handoffs. The signal to candidates: this is a company that picked its problem because the founders lived inside it, not because it was the largest market in healthcare IT.
There's also a clear operational tilt. Tennr's products are built to plug into the existing mess (handwritten orders, faxes, Epic-generated PDFs) rather than to rip and replace workflows. Holterman told Healthcare Brew: "We have an obligation to serve [a] patient if [an] order is handwritten, if it's been typed in a fax, or if it's been electronically generated out of Epic. You really haven't solved something if you can just do it out of an electronic transmission." The hiring message: candidates should expect to spend real time with healthcare operations staff and billing-criteria experts, not just model trainers.
The current openings on Zero G Talent's board reinforce the profile. Eighteen salaried roles are posted in Tennr's New York City office, with a median salary band of $175,000 and a range running $95,000 to $215,000. The salary ladder compresses tightly at the senior tier, which reads as a deliberate signal that Tennr prices seniority across functions similarly rather than letting, say, engineering outrun go-to-market roles.
| Role | Band |
|---|---|
| Enterprise Solutions Consultant | $205,000–$235,000 |
| Senior Backend Software Engineer | $190,000–$215,000 |
| Senior Product Engineer | $190,000–$215,000 |
| Senior UX Engineer | $190,000–$215,000 |
| Enterprise Account Partner | $190,000–$215,000 |
| Product Engineer | $165,000–$190,000 |
The mix tells you what gets hired through: senior individual contributors who can own a slice of the stack, plus customer-facing roles that can translate between referral coordinators and an ML product team. Generalists need not apply.
What current and former employees say, reviews and public accounts, two-sided
No third-party employee-review data, Glassdoor commentary, Blind threads, or former-employee blog posts about Tennr were surfaced in the research for this section. That absence is itself worth flagging: any characterization below is built only from the first-party job-board listings on Zero G Talent, which describe the company and its roles in employer-authored language, closer to recruiting copy than to employee voice. Readers looking for unfiltered current-and-former-employee reviews should treat this section as a map of what's publicly posted, not as a verdict.
What the board listings do offer is a consistent picture of how Tennr positions itself to candidates. The titles themselves tell a story reviewers elsewhere often call out at fast-moving startups: a heavy concentration of "Senior" roles for a company of Tennr's stage, paired with one open Product Engineer seat at the junior end. The Senior UX Engineer posting is notable on its own; design-engineering hybrids at that level usually signal small product teams where one person owns a surface end-to-end. Combined with the Enterprise Solutions Consultant and Enterprise Account Partner roles, the board suggests a workforce built around people who carry accounts or product areas directly, rather than layers of support staff behind them.
The repeated "New York City Office" tag on every posting is the clearest factual thread. Tennr lists no remote roles and no other office locations in the data surfaced, which constrains what current and former employees could say about flexibility or distributed-team norms.
Because no negative accounts, no positive accounts, and no neutral accounts from named current or former employees appear in the research, any summary of "what people say" would be fabricated. The honest read is narrower: as of the board snapshot, Tennr is hiring in a narrow band, in one city, at a senior level, and the public review landscape for the company, at least in the sources surfaced here, is thin enough that candidates shouldn't rely on third-party review sites alone. A reasonable next step for a candidate is to ask former Tennr employees directly through LinkedIn or to request references during the on-site loop; the data here isn't enough to substitute for those conversations.
Who thrives here and who burns out
A company that has grown from a Stanford student project to processing more than 10 million documents a month across 150 healthcare organizations, in roughly four years, is not a place where anyone gets to coast. Tennr's trajectory — founded in 2021, $101 million Series C in mid-2025, a brand-new ad campaign breaking the same year, according to PR Newswire — is the profile of a team that ships, then ships again, then ships something it has never built before (a fax-breaking campaign; a voice-AI phone layer) on a cadence that punishes indecision.
That pace selects for a recognizable type. The person who tends to do well here is the engineer or operator who treats ambiguity as a working condition, not a problem to escalate. The product surface keeps moving — referral intake, eligibility, billing, patient comms, auth review, and now embedded voice AI for outbound payer and scheduling calls — and the company has publicly committed to accelerating a "Center of Excellence" team of operators "who have lived the operational and administrative problems that we deal with," per Holterman. In practice, that means domain hires (former provider-side staff, clinicians-turned-operators, healthcare IT veterans) sit alongside the Stanford engineering core and are expected to push product decisions, not just describe them.
The work environment appears to reward three habits in particular. First, comfort with vertical depth: Tennr trained its proprietary vision-language model, RaeLM, on that same scale of data — 2.3 billion fields and 8,000 payer-criteria sets — and the team's pitch is that it understands clinical notes, scanned forms and checkboxes "unlike generic large language models." Staff who can dig into messy payer criteria and surface where submissions will fail are the ones feeding that engine. Second, bias toward the messy channel. The CEO's own framing in a Fierce Healthcare interview, "we had to bring in voice, because that's where the bottleneck for most of our customers had moved," telegraphs a culture that follows the friction rather than the trend. Third, tolerance for a flat structure where decisions get made close to the work. Repeated product expansions (eligibility, communications, auth review) inside a headcount that still spans roughly 18 salaried roles on the board don't happen through committees.
The inverse profile also reads clearly, even without naming individual reviewers. Anyone who needs a defined lane, a slow planning cadence, or separation between "my job" and "the company's job" is going to feel the squeeze. The same flatness that lets an engineer own a workflow end-to-end is the same flatness that puts a product launch, an ad campaign and a payer integration on the same sprint. The shift from a single intake-and-documentation product to six product lines, plus voice AI, plus a first-ever brand campaign, in roughly a year, is the operational signature of a company that does not buffer its people from its own growth.
The intensity shows up in the shape of the work itself. Front-office customers expect Tennr's system to flag missing insurance information within two minutes of an order arriving, Holterman told Fierce Healthcare, and the voice layer has to hold up against real payer phone trees. That latency and accuracy bar translates, inside the company, into on-call pressure and a fast iteration loop on models and prompts. Employees who want to unplug cleanly between cycles, or who want six-month roadmaps instead of six-week ones, will read that environment as unsustainable rather than exciting.
One tension the public reporting leaves unresolved: Tennr's leadership talks about embedding operator voices and provider expertise into product decisions, which suggests an attempt to keep the human layer from being flattened by velocity. Whether that cushion actually absorbs the pace, or merely decorates it, is a question only current staff can answer, and the published record does not give enough signal to call it.
If you're weighing an offer, the honest filter is this: can you ship into production on a problem you didn't fully understand two weeks ago, in a domain (payer rules, clinical documentation, phone-based intake) that is genuinely weird, without waiting for someone to define your scope? If yes, the broad salary bands and the equity at a Series C company scaling past 150 provider customers make the trade legible. If the answer is no, the same flatness and intensity that make Tennr move fast will make your weeks feel long. The voicemail a patient never has to leave, because the referral was already cleared before they reached the parking lot, is the product, but it is also the pace, and the two are not separable.
Working in frontier tech? Zero G Talent tracks the openings: see every open Tennr role, browse frontier tech jobs, the companies hiring, and the people building the field.