VoiceOS Has Two Live Roles. TheAntiJobBoard Lists None.

The Open Roles: What the Record Shows

A voice-AI startup backed by Y Combinator lists two open roles on its own jobs page. A third-party directory insists it has none. The discrepancy isn't a clerical error — it's the first test applicants face, and it reveals the real screen: at frontier-tech companies where directories disagree on basic facts, the early filter isn't credential-matching. It's signal-verification and founder-fit.

VoiceOS, a Y Combinator company building "the next generation of voice intelligence," shows two active postings on YC's jobs platform: a Customer Support & Discord Community Associate role, open to new graduates and location-agnostic, and a Part-Time Videographer & Video Editor position based in San Francisco. Zero G Talent's board mirrors both listings, with the role appearing as Customer Support & Discord Community Associate (Remote), and the video role as Part-Time Videographer & Video Editor (San Francisco, CA, US). The board shows one role added in the past seven days. Its salary band for salaried positions at VoiceOS is detailed in the table below across a single salaried role.

Yet TheAntiJobBoard, a directory that aggregates startup postings, reports "No open roles today" for VoiceOS as of October 20, 2026. Two live postings on the company's own YC page and a partner board versus zero on an aggregator — the contradiction forces anyone considering an application to check the primary source: VoiceOS on YC's site and the company's own careers page if one exists.

The two roles reveal a pre-seed team shaping its external face. The Customer Support & Discord Community Associate sits at the intersection of user operations and community building; at a voice-AI startup, that means triaging developer feedback, managing a Discord server where early adopters stress-test APIs, and translating friction into product tickets. The compensation aligns with a part-time or early-career contractor arrangement rather than a full-time salaried role. The videographer role, explicitly part-time at roughly 20 hours per week and anchored in San Francisco, signals a founder-led push to produce founder-facing content: demo videos, launch assets, investor updates. Both roles are contractor-grade, not equity-heavy founding-engineer slots. That matters for how the screen works.

Four founders are listed on the company page: Jonah Daian, Ari Ramsan, Kai Brokering, and Parker Bjur. A four-person founding team at YC stage with two contractor openings suggests a product still in the "show, don't tell" phase, building trust with developers and investors through polished demos and responsive community support rather than scaling a sales motion. The conflict between directories doesn't resolve itself. It persists as a signal: the company's public footprint is small enough that aggregators disagree on its existence. Applicants who treat that as noise miss the actual screen.

One Company, Three Descriptions: The Public-Record Conflict

Pull up VoiceOS on Y Combinator and you get a consumer productivity play: "The Siri for Productivity," a universal voice interface that works across macOS and Windows apps (Gmail, Slack, Notion, Cursor), turning speech into polished text and executed actions. The YC page, mirrored by startupintros.com and the company's own LinkedIn, positions VoiceOS as a desktop voice operating system for knowledge workers, founded in 2023 by Daian and Brokering, now in the Spring 2025 batch with a team of two and $500K seed raised in June 2025. Their LinkedIn specialties list "Voice-to-Action, AI Agents, Dictation, Workflow Automation, and Voice operating system." The product site, voiceos.com, shows Agent Mode and Dictation Mode for Mac and Windows. Product Hunt coverage from August 2026 calls it "the universal voice → action for your computer" and notes a #2 Product of the Day finish. The founders' own posts describe 30,000+ users, a hackathon with 150 builders, and a roadmap toward "a real-life JARVIS for your entire computer."

Open oneroadmap.io and VoiceOS is something else entirely: "a phone-first omnichannel communications platform for Main Street and field services businesses" that helps "owners, office managers, dispatchers, and crews manage calls, texts, emails, web chats, vendor messages, and partner referrals in one shared operating record." No mention of desktop dictation, no Mac/Windows clients, no consumer productivity angle. This is B2B SaaS for plumbers, HVAC crews, and local service shops.

Visit voiceosai.com and a third identity appears: "VoiceOS AI - AI Phone Agents for Every Business," a "complete AI front desk: answering, qualifying, booking, and following up, around the clock." Every call answered on the first ring. Custom questions filter prospects. Calendar sync books jobs and sends confirmation texts. This is an AI receptionist product, not a desktop voice OS and not a Main Street CRM.

A fourth variant lives at otomeson.com: "VoiceOS | Complete Agentic Voice System for Enterprise" that "recognizes who's calling, remembers their history and context, and executes tasks end-to-end." Enterprise agentic voice system. No desktop, no Main Street, no front-desk booking — just enterprise call automation.

Four domains. Four mutually exclusive product categories. Consumer desktop voice OS. Main Street omnichannel CRM. SMB AI phone agent. Enterprise agentic voice platform. The YC and LinkedIn profiles agree on founders Daian and Brokering, a two-person team, and a Spring 2025 batch. But Zero G Talent's board lists four founders (the four founders) and shows two open roles. The board's salary band for the single salaried role sits at $73K–$83K median. LinkedIn shows four people at the company including Aidan Mui and QA engineer Veronika Iatckaia. YC says team size two. The discrepancy isn't a rounding error; it's a factor of two to four depending on which directory you trust.

The conflict extends to traction. YC-linked sources cite 30,000+ users and a Product Hunt launch. The Main Street and AI phone agent sites show no user numbers, no Product Hunt badges, no hackathon photos. They read like landing pages for different products that happen to share a name. Whether VoiceOS is pivoting, running parallel experiments, or suffering from name collision with another entity, the public record does not say. What it does say is that an applicant cannot rely on any single directory to understand what this company builds, who it serves, or how many people actually work there. The signal an applicant sends by demonstrating they've reconciled these contradictions (or at minimum, spotted them) may be the only credential that matters at this stage.

Past the Screen: What a 2-Person YC Voice Startup Screens For

VoiceOS lists two founders on its Y Combinator profile (Daian and Brokering) and a team size of two. The Zero G Talent board shows four founder profiles. That discrepancy alone tells you the cap table and the org chart are still fluid. At a pre-seed YC company, the screen isn't a rubric; it's the founders' own filter for the next person who can reduce their personal workload without adding management overhead.

Daian has spent roughly seven years on voice AI. His timeline reads like a vertical slice of the field: audio processing models for voice overs at EzGlobe (2019), end-to-end automated voice translation at Interpret Video (2021), ZKP authentication for voice at Rowzzy (2022), the first AI receptionist on the App Store with NuPhone (2024), and a voice AI interviewer deployed at the largest EV manufacturer. Brokering's path runs through Rakuten engineering and a NASA internship before co-founding NuPhone with Daian. Both founders describe the current product as "solving every part of the voice agent pipeline" with specific emphasis on speech-to-text as "the most neglected part in Voice AI," a line that appears verbatim on the YC Work at a Startup posting.

That technical focus shapes the hire. A two-person team building the interface that works across Slack, Gmail, Notion, VS Code, ChatGPT, Figma, Zoom, Teams, iMessage, and WhatsApp with zero setup needs engineers who can ship across the full stack: audio capture, on-device ASR, intent parsing, cross-app injection, and the new Agent Mode that executes multi-step workflows from a single voice command. The hackathon VoiceOS organized in Tokyo (three hours of keyboardless development using Agent Mode to automate tasks like "Open the customer data and record the latest weight") doubles as a recruiting signal. Sponsors Torema (fitness DX) and Lily Giken (AI development, active Claude Code users) are evaluating voice as a next-generation development methodology. Candidates who show up to that event, or who can articulate why voice-x-coding is a turning point, signal the founder-level adaptability the team actually screens for.

The open roles reinforce this. The associate role remote isn't a traditional support hire; it's a feedback loop into product. The videographer role in San Francisco serves the launch narrative, as Daian's launch video and the "JARVIS" framing on LinkedIn both treat demo quality as a product feature. Neither role appears on a standard career ladder. Both exist because the founders need specific leverage points right now: community signal and visual proof that the product works.

YC's Spring 2025 batch partner is David Lieb. That matters. VoiceOS positions voice as "connective tissue" across existing interfaces — not a replacement for keyboard, mouse, or touchscreen. The product philosophy page repeats: "We do not believe voice replaces every interface. We believe it becomes the connective tissue across them." A candidate who treats voice as a drop-in dictation tool misses the architecture. The screen selects for people who understand that the hard problem is context carryover across apps (the "shared layer of meaning" the founders describe) and who can build toward AGI-era workflows where the OS prepares work, shows what will happen, and asks before executing.

According to Zero G Talent's board data, the board's salary band for the single salaried role sits at $73K–$83K median. That band suggests the filter self-selects for equity alignment and trajectory over cash. The founders have already taken multiple swings at voice products; they're not hiring to de-risk. They're hiring to extend the surface area of what two people can ship before the next fundraise.

The founders' public writing emphasizes that neglected part in Voice AI, Agent Mode's "prepare, show, ask" loop for hands-free environments, and cross-app context carryover as the hard problem. Applicants who engage with these specifics signal that adaptability. The directories disagree on headcount and role count. The founders don't. They screen for the person who verifies the product, not the posting.

Who Should Apply, and Who Should Look Elsewhere

VoiceOS is a two-person, pre-seed voice-AI startup in San Francisco (Team Size: 2 per its Y Combinator profile (Batch P25, founded 2023)) and the first-party board data shows exactly one salaried role on file with a band (median) per the table above. The two live openings as of this week are a the videographer role (San Francisco, on-site shoots required) and a the associate role (Remote). That is the entire hiring surface. Applicants who match should be comfortable with a founder-led screen: the four founders run sales, product, and payroll themselves; there is no recruiting team, no ATS, and the Anti Job Board notes that "your application lands with a founder who is also running sales, product and payroll." Cold outreach to those founders outperforms the application form consistently.

The videographer role demands professional gear (camera, gimbal, lighting, audio), advanced editing in Premiere/Final Cut/DaVinci, fluency in TikTok/Reels/Shorts pacing, and the ability to shoot on location in San Francisco; "Based in San Francisco (on-site required)" is explicit. The customer-support role is remote but expects Discord community management and availability for a consumer voice product ("The Siri for Productivity" per the Anti Job Board). Both roles require "relevant experience and ability to operate autonomously" at a <50-person consumer company where "broad remit, direct access to founders, and equity that still means something if the company works out" is the trade-off. The hiring loop is three stages over roughly seven days; roles at this stage typically draw 50–100 applicants in the first two weeks, and applying within 72 hours of posting is the single biggest lever a candidate controls.

Who should look elsewhere: anyone targeting senior leadership — there is no layer above the founders. Remote-only candidates for the videographer role (on-site is non-negotiable). Non-voice-AI specialists — the product is a voice interface for macOS/Windows, not a general LLM wrapper. Candidates expecting structured onboarding, defined career ladders, or a recruiting coordinator — the obstacle "isn't a queue or an ATS, it's being seen at all." Salary seekers above the median for salaried work or the contractor ceiling per the table above. Applicants who recycle CVs — "A recycled CV gets rejected fast at VoiceOS (<50 people). They notice." And anyone unwilling to verify the basics themselves: directories disagree on headcount (YC says 2, Anti Job Board says <50), role count (Anti Job Board shows "No open roles today" while the board shows two live), and even product description ("consumer voice OS" vs. "YC computer-automation tool" vs. "Main Street phone platform"). The early screen is not credential-matching; it is signal-verification and founder-fit. If you cannot confirm the company's actual size, open roles, and product focus from primary sources before you write the first line of your outreach, you will not clear the founders' inbox.

Verify Before You Apply: A Hiring Checklist for Frontier Tech

The directories disagree on VoiceOS. One calls it a consumer voice operating system. Another lists it as a Y Combinator computer-automation tool. A third describes a Main Street phone platform for small businesses. The headcount ranges from two to over 20. The open-role count swings from zero to three. When the public record cannot agree on what a company is, the burden shifts to the applicant: verify before you tailor a single bullet point.

Start with the careers page. Type the URL yourself; do not click a link from a job board, a recruiter email, or a LinkedIn post. VoiceOS's board page shows the two live openings: a the associate role (remote) and a the videographer role (San Francisco). If a third role appears on Indeed or Wellfound but not on the company page, treat it as unconfirmed. The Federal Trade Commission recorded $501 million in US job-scam losses for 2024, up from $90 million in 2020. The median victim lost roughly $1,500. Forty percent of 2024 reports involved "task" or "gamified" scams, the fastest-growing category. Fourteen million-plus scam postings circulate annually. Sixty-eight percent target remote roles. VoiceOS's community-associate role is remote. That alone warrants extra scrutiny.

Search the company name plus "scam." Check Reddit, Hacker News, and the FTC's complaint database. If VoiceOS has ever been impersonated, you will find threads. Next, verify the recruiters. The founders (the four founders) should have LinkedIn profiles with consistent histories, mutual connections, and Y Combinator affiliation. A recruiter messaging you from a generic Gmail address or a domain that does not match voiceos.ai is a red flag. So is any request to move the conversation to Telegram, WhatsApp, or Signal before a video interview. Upfront fees, crypto payments, gift cards, or requests for bank details prior to an offer letter are not "standard process"; they are theft.

Run the job description through a scam-pattern analyzer. VerifyJobs.org offers a free engine (50+ rule-based indicators plus an XGBoost/BERT model, 0.819 AUC, 1.00 recall) and a Chrome extension that scans pasted text or the active tab. Real Job Check cross-references the posting against the careers page, the sender's domain, realistic pay bands, and known scam patterns. Zero G Talent's board salary band runs per the table above for salaried roles; the community-associate role pays, as Zero G Talent's figures put it, per the table above and the videographer role per the table above. Both fall within plausible ranges for early-stage, part-time, or contractor work.

Confirm the Y Combinator connection. VoiceOS appeared in the YC job board's "all roles" feed, which lists many YC startups actively hiring. But YC's directory does not vet every posting for accuracy; it aggregates. Check the YC company directory directly for VoiceOS's batch, batchmates, and demo-day materials. A pre-seed voice-AI startup with two to four founders typically screens for founder-level adaptability. The screen is not credential-matching; it is signal-verification — can you demonstrate you have already done the work the role requires, in public, with code or recordings you can link?

Finally, cross-reference the headcount. If LinkedIn shows over 20 employees but the founders list four and the board shows two active roles, ask in the first conversation: "Who are the other employees, and what do they do?" A legitimate pre-seed team will answer directly. A fabricated one will deflect. The checklist is not paranoia. It is the new baseline for frontier-tech hiring — where the directories conflict, the applicant who verifies wins.

Working in frontier tech? Zero G Talent tracks the openings: see every open VoiceOS role, browse frontier tech jobs, the companies hiring, and the people building the field.