VerifyHire AI: Smart JOB

Legitimacy System

Project Details Team Members

Entrepreneurship Name Enrollment Number
Subject
Mindset
Aditya Goel 00116401523
Subject Code MS307
Ishant 01616401523

Executive Summary
Online job fraud is a rapidly growing threat. In fact, reported losses from online job
scams tripled between 2020 and 2023ftc.gov. Many of these scams are disguised as
fake job postings engineered to steal personal information or moneyindeed.com.
verifyHire is a proposed SaaS platform that authenticates job listings before they reach
applicants. By cross-checking each listing against company databases, official
websites and reputable sources, the platform would grant a “verified” seal only to
legitimate postings. This business plan outlines how verifyHire will fill a clear gap in the
recruiting market: protecting job seekers from fraud and restoring trust in online hiring.

Problem Statement
Fraudulent job ads are pervasive and damaging. In a recent Aerotek survey, 79% of
job-seekers reported encountering suspicious postings at least some of the
timeaerotek.com, and 64% encountered obvious fraud, with nearly 25% admitting
they’d been victimized by a scamaerotek.com. The Better Business Bureau estimates
that 14 million people are exposed to employment scams each year, leading to over
$2 billion in aggregate lossesbbbmarketplacetrust.org. FTC data mirror these trends:
reported consumer losses from job scams exceeded $220 million in just the first half of
2024ftc.gov. Most fake job ads aim to harvest personal data or money from
applicantsindeed.com, exploiting vulnerabilities in the hiring process. Even major job
sites admit that, despite screening measures, scammers can slip fraudulent listings onto
their boardsindeed.com. The result is wasted time, financial harm, and growing distrust
among job-seekers – a problem that demands a technical solution for verifying listing
authenticity before applicants are harmed.

VerifyHire AI: Smart JOB Legitimacy System 1

 Business Opportunity
The market for online recruiting solutions is large and growing rapidly. Fortune Business
Insights projects the global online recruitment technology market to expand from
$13.2 billion (2024) to $37.8 billion (2032)fortunebusinessinsights.com. In the U.S.
alone, that sector reached about $5.2 billion (2024), with a forecast to $15.5 billion by
2032psmarketresearch.com. Job-platform usage is enormous – for example, LinkedIn
boasts ~1.3 billion members worldwide and over 71 million companies
listednews.linkedin.comnews.linkedin.com; users submit on the order of 9,500 job
applications per minutenews.linkedin.com. This scale creates a vast pool of potential
scam targets. At the same time, hiring managers are increasingly adopting AI and
automated tools to handle large applicant volumespsmarketresearch.com. The trend
toward tech-enabled hiring suggests a strong appetite for new solutions. Yet none of
the leading platforms (LinkedIn, Indeed, Monster, etc.) currently offer a comprehensive
verification service for every posting. verifyHire can exploit this niche by integrating
trust and authenticity checks into the recruiting tech stack, positioning itself as a
complementary service in the booming HR technology ecosystem.

Proposed Business / Product Description

verifyHire will be a software-as-a-service (SaaS) platform dedicated to verifying job
ads. It will ingest postings from job boards, company career sites, and social media,
then apply automated validation steps. For each listing, verifyHire will cross-reference
the employer’s official registries, domain information, and public profiles (e.g. LinkedIn
company pages) to confirm legitimacy. Listings that pass all checks receive a visible
“Verified” badge; suspicious ones are flagged for review or hidden. Employers can opt
into the system via API or plug-ins: they might subscribe to get each of their ads vetted,
or pay per-posting verification. Job-seekers accessing verifyHire’s interface (or a
partner job board) will see only vetted opportunities marked safe.
This approach formalizes best practices already noted by recruiters. For example,
staffing firm Aerotek states that they “implement various strategies to verify the
authenticity of our job listings”aerotek.com as part of their candidate care. verifyHire
automates and scales this concept. We will leverage modern technologies (potentially
including blockchain-based ledgers as emerging in recruitment
techpsmarketresearch.com) to create an immutable audit trail for each posting. In
practice, verifyHire serves both B2B clients (employers and job boards seeking to
protect their brand) and B2C users (job-seekers looking for safe listings). Additional
features may include crowdsourced flagging (where candidates report suspect ads) and
integrations with background-check or identity-verification services. Overall, the
product is a “trust layer” for online hiring that all stakeholders can rely on.

Vision & Mission Statements

VerifyHire AI: Smart JOB Legitimacy System 2

 Vision: Every job-seeker can browse listings with confidence that each opportunity is
real and legitimate. (A trustworthy job market where fraud is eliminated.)

Mission: To empower job-seekers and employers by verifying every job listing’s

authenticity using advanced technology and data sources. (Providing a seal of trust on
genuine job opportunities and safeguarding people’s job searches.)

Market Analysis
The addressable market includes online job platforms, staffing agencies, and
employers who post jobs. Globally there are thousands of job boards and recruitment
sites. Key segments include: (1) large publicly-listed boards (Indeed, LinkedIn,
ZipRecruiter), (2) vertical niche boards (engineering, healthcare, etc.), (3) corporate
career sites for mid-to-large companies, and (4) staffing/recruitment firms’ client
postings. With 1.3B+ professionals and 71M companies on LinkedIn
alonenews.linkedin.comnews.linkedin.com, and millions of job ads active at any time, the
scale is massive. The rise of remote and hybrid work (35% of U.S. employees in 2023
worked remotelypsmarketresearch.com) has expanded the geographic reach of listings
– any scammer worldwide can now target local applicants via global boards.

Major industry trends support our value proposition. Hiring tech is embracing AI and
automation, from résumé-screening to chatbotspsmarketresearch.com. Similarly,
verification via technology is an emerging area (for example, blockchain-based
recruitment solutions are cited as a growth segmentpsmarketresearch.com). Job
seekers are also demanding safer experiences: social media has seen groups exposing
“ghost jobs” and fake listings. Regulatory attention (new laws on ghost job
transparency) further underscores the market’s focus on honesty. In sum, the
combination of a large, growing recruitment market and rising demand for trust creates
a strong opportunity for verifyHire.

Business Model
verifyHire will operate on a B2B SaaS subscription basis, complemented by
transactional fees. Key revenue streams may include:

Employer/Board Subscriptions: Organizations (job boards, recruiters, corporate HR)

subscribe to have all their postings verified. For example, a mid-sized company
might pay an annual fee for unlimited verifications or a per-post pricing (e.g. $10–
$20 per listing).

Verification Fees: Alternatively, we may charge on a per-verification basis. A

company posting 100 jobs a year might pay a bulk rate for verifying each.

Value-Added Services: Premium analytics or integrations (e.g. detailed audit logs,

compliance reports, or linking to background check partners) could be added
services for extra fee.

VerifyHire AI: Smart JOB Legitimacy System 3

 Job-Seeker Premiums: The core job-search function for individuals would be free.
We could introduce optional premium features (e.g. personalized fraud alerts, career
coaching, or prioritized alerts for verified roles) as micro-transactions or
subscription upsells.

Ad/Referral Partnerships: With compliance and high trust, we could offer non-
intrusive advertising or earn referral fees from trusted career services.

This mixed model ensures diversified income. The primary focus will be enterprise
buyers (who value risk reduction) rather than direct job-seeker sales, making verifyHire
financially attractive as a B2B tool with B2C touches.

Marketing Strategy
To reach customers, verifyHire will pursue a two-pronged approach targeting
employers/boards and job-seekers:

Partnerships with Job Platforms: We will engage major job boards and HR software
providers to integrate verifyHire. Demonstrating that our technology reduces fraud
can make us a value-add or premium feature for them. Attending HR tech
conferences and networking with industry associations (SHRM, HR Tech, etc.) will
seed these collaborations.

Direct Sales to Enterprises: A small sales team will target large employers and
recruitment agencies that are likely to value brand protection. Case studies showing
reduced fraud can persuade HR departments and compliance officers to buy in.

Digital Content & Social Proof: We will publish expert content (blogs, webinars) on
avoiding job scams to build credibility. Outreach on LinkedIn and professional
forums highlighting our verification checks can attract job-seekers and HR
influencers alike. User testimonials and reported success (e.g. “This listing saved me
from a scam”) will be highlighted in marketing collateral.

Search/SEO and PR: Since job-seekers search tips on spotting scams, SEO content
on “verify job listing” can drive organic traffic to our service. Press releases when
we achieve milestones (seed funding, partnerships) will raise brand awareness.

This strategy combines inbound (content/SEO) and outbound (B2B sales, partnerships)
tactics. We will measure success by pilot program signups, conversion of free users to
paid integrations, and customer retention rates.

Operations Plan
Team & Development: We will build a lean core team initially: software engineers (to
develop the verification engine and UI), data specialists (to access and interpret public
company records), and a few compliance/legal advisors (to navigate data privacy).
Development will follow agile iterations, starting with an MVP that can vet listings from

VerifyHire AI: Smart JOB Legitimacy System 4

 one major job board. We will use cloud infrastructure (AWS/GCP) to scale compute for
automated checks. The tech stack may include machine learning (for pattern detection
in ads) and API integrations with business registries.
Data & Verification Sources: A key operation is maintaining reliable data feeds. We will
integrate with commercial business data providers and use web crawlers/scrapers for
public info (e.g. checking a company’s real website, registration number). Over time, we
plan partnerships with identity verification and background-check vendors to validate
company contact info when needed.
Customer Support & Review: In addition to automation, some human review will be
necessary to handle edge cases or appeals. We will train a small verification team to vet
listings that algorithms flag as uncertain. Customer support channels (email, chat) will
assist employers whose jobs were flagged incorrectly.

Partnerships: We will pursue alliances (e.g. industry alliances, developer programs)

that allow easier API access to job boards. Aligning with security/chamber of commerce
groups may enhance legitimacy. We will also monitor and adapt to legal requirements
(e.g. legislative changes on job-post disclosures) so our operations remain compliant.

Financial Plan (Realistic Projection)

We present a three-year financial outlook under conservative assumptions. Key
assumptions include moderate market penetration and typical SaaS economics:

Year 1: Focus on product development and pilot customers. Expenses ($500K) for
development, data acquisition, and marketing will exceed income. We project
minimal revenue (~$50K) from a handful of pilot clients (covering some pilot costs).

Year 2: Launch full platform to early adopters. Assume we secure ~5 enterprise

clients (subscription plus verification fees), averaging $50K annual contracts each.
That yields roughly $250K revenue. Marketing and staffing costs grow ($400K), but
R&D tapers. Net loss is expected but narrowing (–$150K) as revenue scales.

Year 3: Expansion phase. If verifyHire captures more of the market, say 20 midsize-
to-large clients by year-end (totaling ~$1M in recurring revenue), with additional fee-
for-service income (~$200K), total revenue could reach $1.2M. Operating costs may
be ~$800K. This could result in a small profit.

Conservatively, break-even is achieved late Year 3 or Year 4. These figures assume a

verification fee around $15–$20/listing and moderate growth. We would refine these
projections with detailed customer pipelines and feedback. Additional funding (e.g.
seed investment) would cover early losses. Overall, the plan anticipates steady growth
in ARR and eventual self-sustainability.

Risk Analysis and Mitigation

VerifyHire AI: Smart JOB Legitimacy System 5

 Key risks and our countermeasures include:

Technological Risk: Verification algorithms may miss clever scams (false negatives)
or flag legitimate posts (false positives). Mitigation: Continuous machine learning
model training with real-world examples, plus human review for flagged cases. We
will iterate on logic (e.g. NLP detection of scammy language) to improve accuracy.

Market Adoption Risk: Major job boards or employers may be slow to adopt an
external verifier, preferring to manage fraud internally. Mitigation: We will target
early adopters in privacy/security-minded industries and emphasize partnerships.
Offering a white-label or integration option can lower adoption friction. A freemium
tier can let job-seekers drive viral demand (if users demand “verified” tags on
listings).

Competitive Risk: Large platforms (LinkedIn, Google Jobs) might build their own
verification features. Mitigation: verifyHire aims for flexibility by working with any
platform. Early patents on novel verification methods could create barriers. If a
competitor emerges, our advantage would be our specialized focus and first-mover
data.

Regulatory Risk: New laws could change job-posting rules. For example, states like
Kentucky and California are proposing (or passing) laws requiring employers to
disclose whether a posting is an actual vacancycongress.govcongress.gov. While
such regulation reflects the ghost-job problem, it could impose compliance burdens.
Mitigation: verifyHire can build in compliance tools (e.g. automated posting status
flags) as value-added features, helping clients meet legal requirements. We will
maintain close legal monitoring (for example, Ontario’s upcoming rules on
disclosurecongress.gov) and ensure our service adapts quickly.

Data Privacy/Security Risk: Handling job data and possibly personal data (like email
domains) involves privacy regulations (GDPR, etc.). Mitigation: We will use only
business-public data and anonymize user logs. Compliance frameworks (e.g. SOC
2) will guide our data policies.

Overall, we will manage risk by technical diligence, legal awareness, and by aligning our
value proposition with both market demand and regulatory trends. In fact, having clear
verification can help companies avoid fines or reputational damage, turning a potential
regulatory risk into an opportunity for our service.

Conclusion
The verifyHire platform addresses a pressing and growing need in the recruitment
industry. By preventing fake job postings from harming candidates, we create value for
both job-seekers and employers. The business opportunity is supported by large
market size and fast growth in hiring
technologyfortunebusinessinsights.compsmarketresearch.com. Our solution leverages

VerifyHire AI: Smart JOB Legitimacy System 6

 current tech trends (AI, blockchain, automation) and fills a gap no existing competitor
covers. With a clear go-to-market strategy and prudent financial planning, verifyHire
has the potential to become the standard “fraud filter” in online hiring. By restoring
trust, we improve outcomes for job seekers, enhance employer reputation, and capture
a valuable niche in the HR tech ecosystem.

References
Federal Trade Commission (2024). New FTC Data Show Skyrocketing Consumer
Reports About Game-Like Online Job Scams (Press Release, Dec 12, 2024).

Aerotek (2023). Job Scams on the Rise: How to Spot and Avoid Fraud in Your Job
Search.

BBB Institute for Marketplace Trust (2020). Protecting Yourself from Employment
Scams.

Fortune Business Insights (2023). Online Recruitment Technology Market Size,

Share & Industry Analysis (2023–2032).

P&S Intelligence (2024). U.S. Online Recruitment Services Market Size & Forecast.

Indeed (2025). 17 Common Job Scams and How to Protect Yourself (Career Advice
article).

Congressional Research Service (2024). “Ghost” Job Postings (Issue Brief, Library
of Congress).

LinkedIn Newsroom (2025). About Us – Statistics.

Product Requirements Document (PRD)

1. Problem, Goals, and Non‑Goals
Problem: Rampant fake job listings erode trust, waste time, and cause financial loss
for candidates and reputational harm for employers.

Primary goal: Accurately verify job postings at scale and surface a clear, portable
trust signal (“Verified” badge + machine‑readable assertion).

Secondary goals: Provide explainable verification outcomes and an auditable trail.

Offer APIs and integrations that minimize adoption friction.

Non‑goals: We are not a background‑check or candidate‑screening company. We do

not guarantee employer conduct post‑hire.

2. Success Metrics and KPIs

VerifyHire AI: Smart JOB Legitimacy System 7

 Precision/recall of verification outcomes on labeled datasets.

Reduction in scam exposure rate for partner platforms.

Time‑to‑verification (p50, p95) per listing.

“Lift” in application rate and conversion on verified postings.

Publisher adoption: number of integrated job boards and career sites.

Enterprise ARR, net revenue retention, and payback period.

False positive rate for legitimate posts and appeal cycle time.

3. User Personas
Job‑seeker: Needs quick, trustworthy signals; prefers simple badges and clear
reasons.

Recruiter/TA manager: Wants low‑friction bulk verification, audit logs for compliance,
and SLA guarantees.

Job board/platform PM: Seeks fraud reduction, stronger brand trust, minimal latency
impact, and easy API/webhook integration.

Compliance/security lead: Requires policy mappings (GDPR, SOC 2), data‑handling

controls, and vendor security documentation.

4. User Stories
As a job‑seeker, I want to see a clear badge and reasons so I can decide safely in
seconds.

As a recruiter, I want CSV/API bulk verification so my postings publish automatically

when cleared.

As a platform PM, I want webhooks to update listing states and detailed error codes
when checks fail.

As compliance, I want immutable evidence for audits, with least‑privilege access and
redaction.

5. Scope and Constraints

Start with English‑language postings; expand to multilingual in phase 2.

Target latency: under 5s for 80% of posts, with async deep checks for the
remainder.

Data: Prefer authoritative sources; fall back to heuristics with confidence scores.

Privacy: Personal data minimization. No storage of candidate PII.

VerifyHire AI: Smart JOB Legitimacy System 8

 System Architecture
High‑Level Components
Ingestion layer: APIs, file import, partner connectors, RSS/sitemaps, and scraping
where permitted.

Verification engine: Multi‑stage pipeline with deterministic checks, ML risk scoring,

and rule evaluation.

Evidence graph: Structured store of entities (employer, domain, job post, contact
channels) and edges with provenance.

Badge service: Issues signed verification assertions (JWS/JWT) and renders

embeddable badges.

Reviewer console: Human‑in‑the‑loop tool for edge cases and appeals.

Integration SDKs: REST, GraphQL, and client libraries; webhooks for state changes.

Data Flow
1) Ingest posting payload → normalize schema.
2) Run deterministic checks (domain, DNS, WHOIS, MX, DMARC, SSL, official links).
3) Query authoritative registries and knowledge sources.
4) Run ML models: language cues, anomaly detection, contact mismatch risk, salary
plausibility.
5) Aggregate signals into a confidence score with explainable reasons.
6) Decide: Verified, Flagged, or Rejected + required remediation.
7) Persist evidence and publish signed badge + webhook.

Technical Stack
Backend: TypeScript/Node or Kotlin, gRPC/REST, OpenAPI; message bus
(Kafka/NATS).

Storage: Postgres for transactional data; Elasticsearch/OpenSearch for search;

graph DB (Neo4j) for evidence relationships; object store (S3/GCS) for artifacts.

ML: Python, Hugging Face, Scikit‑learn, LightGBM; feature store (Feast).

Infra: Kubernetes, Terraform, Helm; observability via OpenTelemetry, Prometheus,

Grafana.

Security: mTLS intra‑service, OIDC for auth, KMS/HSM for keys, Vault for secrets.

VerifyHire AI: Smart JOB Legitimacy System 9

 Verification Methods
Deterministic Checks
Domain control validation: Posting domain vs official employer domain;
SPF/DKIM/DMARC alignment.

Site integrity: TLS certificate CN/SAN match, HSTS, canonical employer links.

Corporate identity: Government/corporate registries, tax IDs where available,

address match with utility or mapping data.

Contact channel validation: Emails on corporate domain, phone in official directory,

LinkedIn company page URL consistency.

Job data consistency: Role, location, salary range coherence with employer history
and industry baselines.

Machine Learning Signals

Text classification: Scam cue phrases, urgency patterns, payment requests,
grammar anomalies, templated layouts.

Representation learning: Employer embeddings, cross‑post similarity, outlier

detection across a platform.

Graph features: Distance to known trusted entities, shared infrastructure with known
bad actors.

Ensemble scoring: Calibrated probabilities with Platt scaling; explainable SHAP

summaries.

Evidence and Explainability

Every check produces a typed Evidence object with source, timestamp, and hash.

Explainable summary shown to users: “Verified because A, B, C; low risk due to X;

no red flags on Y.”

For flags: precise remediation advice, e.g., “Publish jobs via careers.example.com or
add a verified TXT record.”

Trust and Badge Protocol

Assertion Format
JSON Web Signature + JSON‑LD context for interoperability.

VerifyHire AI: Smart JOB Legitimacy System 10

 Claims: job_id, employer_id, verified_at, expiry, checks_passed, confidence,
evidence_hash.

Public keys published via a JWKS endpoint; short‑lived tokens to prevent replay.

Display and Anti‑Abuse

Dynamic badge fetch with cacheable, tamper‑evident image + alt text.

Deep‑link to verification detail page with shareable permalink.

Optional on‑page widget that validates the JWS against our JWKS in the client.

API Design
Core Endpoints
POST /v1/jobs: Create or submit a job for verification.

GET /v1/jobs/{id}: Retrieve status, confidence, and reasons.

POST /v1/jobs:bulk: Bulk submit up to N posts.

POST /v1/appeals: Submit evidence to appeal a decision.

GET /v1/badge/{id}.svg: Retrieve badge asset; headers include signed metadata.

Webhooks: job.verified, job.flagged, job.rejected, job.expiring.

Authentication and AuthZ

OAuth 2.1 client credentials for server‑to‑server.

Fine‑grained scopes: jobs:write, jobs:read, badge:read, evidence:read.

Tenant model with RBAC roles: admin, developer, reviewer, auditor.

Security, Privacy, and Compliance

Data minimization and purpose limitation. No candidate PII.

SOC 2 Type II roadmap; ISO 27001 alignment; DPA and SCCs for EU transfers.

Encryption in transit and at rest; key rotation policies; comprehensive logging with
retention tiers.

Vendor risk management, SDLC security gates, SAST/DAST, SBOM.

Human Review Operations

Queue design: priority by partner SLA, severity, and customer tier.

VerifyHire AI: Smart JOB Legitimacy System 11

 Reviewer console: side‑by‑side evidence, quick templates, redaction tools, and blind
QC sampling.

Quality metrics: inter‑rater agreement, calibration sessions, drift monitoring.

Product Design and UX

Job‑seeker UI: simple badge, one‑click details, “Why verified?” clarity, and
report‑a‑problem.

Employer portal: dashboard with pass rate, reasons, remediation suggestions,

exportable audit.

Platform dashboard: SLA graphs, latency heatmaps, top failure reasons, and
signature health.

Competitive Landscape
Point‑in‑time screening by major boards vs. continuous, cross‑source verification.

Identity vendors focus on person identity, not posting authenticity.

Our moat: evidence graph, signed assertions, and platform‑friendly latency + UX.

Go‑to‑Market Deep Dive

Land with mid‑market ATS/HRIS partners; embed as a premium verification toggle.

Usage‑based pricing aligned to verified post volume with tiered discounts.

Pilot programs with measurable scam reduction KPIs and co‑marketing case studies.

Pricing and Packaging

Starter: $499/mo + $0.30 per verification, 50k API calls, email support.

Growth: $2,000/mo + $0.20 per verification, SSO, priority webhooks, 99.9% SLA.

Enterprise: custom; dedicated VPC peering, audit export, custom evidence

connectors.

Financial Model Details

COGS drivers: data providers, inference, storage, and human review.

Gross margin targets >75% by Year 3 via model efficiency and caching of evidence.

CAC and payback modeled per segment; partner channel reduces blended CAC.

VerifyHire AI: Smart JOB Legitimacy System 12

 Roadmap
Q1: MVP with domain and contact verification + initial ML classifier.

Q2: Evidence graph v1, badge service, and first ATS integration.

Q3: Multilingual support, registry connectors expansion, reviewer console v2.

Q4: Signed assertions v2, enterprise features, and SOC 2 audit prep.

Risks and Mitigations (Expanded)

Adversarial adaptation → Regular red‑team exercises and honeypot postings.

Data outages → Multi‑vendor redundancy and graceful degradation with confidence

bands.

Legal landscape shifts → Feature flags to adapt jurisdictional rules.

Ethics and Fairness

Bias audits for ML models; fairness KPIs across industries and geographies.

Transparent appeals and publisher education to reduce inadvertent penalization of

small firms.

Internationalization and Localization

Language detection, locale‑aware features, registry connectors per market.

Right‑to‑left and script‑specific tokenization for ML.

Accessibility
WCAG 2.2 AA for all surfaces. Keyboard‑first flows and screen reader labels.

Analytics and Observability

Golden signals on ingestion, verification, and badge retrieval.

Feature attribution for model decisions; drift and data quality monitors.

FAQ
What if a legitimate startup has no registry footprint? We support alternative proofs
like DNS TXT verification and founder email verification.

VerifyHire AI: Smart JOB Legitimacy System 13

 Do badges expire? Yes, assertions are short‑lived and auto‑refresh when source
evidence remains valid.

Can platforms hide unverified jobs? We provide policy hooks; platforms decide
enforcement levels.

Glossary
Evidence graph, assertion, confidence, precision/recall, DMARC, DKIM, JWKS.

Appendix A: Example Verification Report

Post ID, Employer ID, Checks Passed, Confidence: 0.92, Reasons: domain alignment,
registry match, contact channel verified.

Appendix B: Data Model Sketch

Entities: Employer, Domain, Job, Contact, Evidence, Assertion, Appeal.

Relationships: Employer—owns→Domain; Job—published_on→Platform; Job—

has→Evidence.

Extended Whitepaper
1. Background and Context
Online recruitment has evolved through three eras: classifieds to portals, portals to
aggregators, and aggregators to platform ecosystems. Each transition scaled volume
and reach, but also broadened the attack surface for scammers. Today’s threat actors
operate like growth hackers, A/B‑testing posts, rotating infrastructure, and exploiting
platform loopholes. A trust layer is overdue.

1.1 Threat Taxonomy

Vanity brand impersonation: Fake postings using a well‑known logo to harvest PII.

Payroll redirect: Offer letters used to switch direct‑deposit details.

Application fee scams: Upfront payment requests for training or equipment.

Phishing via ATS lookalikes: Fake portals mimicking Workday/Greenhouse.

Social engineering: Recruiter lookalikes on LinkedIn with recently created profiles.

Location misdirection: Remote roles with geographic tax or export‑control issues.

1.2 Impact Model

VerifyHire AI: Smart JOB Legitimacy System 14

 Individual: Financial loss, identity theft, career setback, emotional distress.

Employer: Brand damage, candidate trust erosion, support burden, legal exposure.

Platform: Reduced NPS, moderation costs, advertiser concern, regulatory scrutiny.

2. Formal Problem Definition

Given a job post J with attributes {employer E, domain D, contact C, content T, location
L, salary S, platform P}, decide label y ∈{Verified, Flagged, Rejected} and produce a
justification set R with evidence references. Optimize for expected social cost reduction
subject to latency, accuracy, and explainability constraints.

3. Evidence Graph Design

Nodes: Employer, Domain, Certificate, RegistryRecord, JobPost, ContactChannel,
SocialProfile, Platform, IP, ASN.

Edges: owns, resolves_to, publishes, references, verified_by, similar_to,

previously_flagged.

Provenance: Each edge carries source, timestamp, hash, and confidence. Hash
chains allow audit trails.

4. Registry and Data Connectors

Corporate registries: OpenCorporates, national registrars, SEC/EDGAR, MCA (India),
Companies House (UK).

DNS/PKI: DNS, WHOIS/RDAP, CT logs, OCSP, Certificate Transparency datasets.

Communication: Email provider fingerprints, MX/SPF/DKIM/DMARC, Twilio Lookup

for phone.

Social/Professional: LinkedIn company pages, GitHub orgs, Crunchbase, G2 reviews

for legitimacy signals.

Salary/Occupation: O*NET, BLS, ESCO to validate role‑salary coherence.

Geospatial: OpenStreetMap, Google Places, address standardization.

5. Model Development
Datasets: Curated positive set from verified corporate feeds; negative set from
reported scams, honeypots, and platform takedowns.

Features: n‑grams, contextual embeddings, domain age, MX alignment, CT log

issuer score, registry match score, URL path entropy, phone region mismatch.

Training: Stratified splits by employer to prevent leakage, temporal validation,

adversarial sampling.

VerifyHire AI: Smart JOB Legitimacy System 15

 Evaluation: Precision@K, AUROC, AUPRC, selective classification with abstain
option.

Robustness: Back‑translation, typos, unicode confusables, prompt‑injection style

bait in post content.

6. Human‑in‑the‑Loop Protocols
Triage levels: Auto‑clear, reviewer‑required, auto‑reject with appeal.

Playbooks: Domain claim challenges, TXT record verification, corporate email

confirmation.

Quality: Golden set rotation, inter‑rater agreement (Cohen’s κ), reviewer dashboards.

7. Legal and Regulatory Landscape

Fraud and unfair practices: FTC Act, state consumer protection laws.

Data protection: GDPR, CCPA/CPRA, DPDP (India). Our approach: minimal data, DPA,
SCCs.

Employment transparency: Pay range and “active vacancy” disclosures in select

jurisdictions.

Safe badges: Avoid deceptive implication of government endorsement; clear

language in UX.

8. International Expansion Playbook

Phase EU: Companies House, Handelsregister, INSEE Sirene, UBO registers where
accessible.

Phase APAC: MCA (India), ACRA (Singapore), ASIC (Australia), NZBN (New Zealand).

Language: Tokenization per script, transliteration for entity resolution, locale‑aware

name matching.

9. Partner Integrations
ATS: Greenhouse, Lever, Workday, SAP SuccessFactors, Taleo via APIs and
webhooks.

Job boards: Indeed, ZipRecruiter, LinkedIn via partner programs where available.

HRIS: Rippling, BambooHR for employer verification shortcuts (opt‑in).

10. Performance Engineering

Caching: Evidence TTLs by source; negative caching to avoid repeated misses.

VerifyHire AI: Smart JOB Legitimacy System 16

 Degradation: Confidence bands with UI hints when sources are degraded.

Parallelism: Fan‑out evidence fetch with bounded concurrency; circuit breakers per
connector.

11. Security Deep Dive

Key management: HSM‑backed signing keys, JWKS rotation, audit trails.

Supply chain: SBOM, signed images, provenance via SLSA levels.

Abuse: Rate limiting, partner‑key reputations, anomaly detection for bulk submits.

12. Economics and Unit Costs

Per‑verification cost components: data API calls, inference, storage, and reviewer
minutes.

Sensitivity analysis: Price vs margin under low/high review rates and connector
pricing.

Levers: Model precision, batched queries, registrar partnerships, edge caching.

13. Pilot Design

Selection: 2 job boards and 2 mid‑market employers across different industries.

Baseline: Measure pre‑pilot fraud reports, CS tickets, time‑to‑takedown.

Intervention: Enable verification, monitor changes in complaint rates and application

conversion.

Success: 60%+ reduction in scam complaints, ≤2% false positive rate, p95 latency
≤8s.

14. Customer Success and SLAs

Standard SLA: 99.9% monthly badge availability, webhook delivery guarantees,
support response targets.

Incident response: Transparent status page, RCA within 5 business days for Sev‑1.

15. Detailed Roadmap Epics

Evidence graph v1, ML v1, Badge service, Connectors, Reviewer console,
Compliance toolkit, Partner SDKs.

Milestones with acceptance criteria and owner roles.

VerifyHire AI: Smart JOB Legitimacy System 17

 UX and Content Specifications
1. Badge States
Verified: Green badge with check icon, tooltip “Posting authenticity verified by
verifyHire.”

Under review: Amber badge, tooltip “Additional checks in progress.”

Not verified: Gray badge, tooltip “Publisher has not verified this posting.”

Rejected: Red badge, tooltip with brief reason.

2. Copy Guidelines
Avoid implying job quality or pay; focus solely on authenticity.

Provide concise reasons: “Domain matches employer DNS,” “Registry record

validated,” “Contact email is corporate.”

3. Accessibility Details
SVGs with descriptive titles, focus states, and high‑contrast modes.

Keyboard navigation and ARIA labels for badge widget.

4. Flows
Employer onboarding: DNS TXT challenge, OAuth app install, initial backfill
verification.

Appeal flow: Submit evidence, track status, receive decision with rationale.

Engineering Implementation Plan

1. Sprints 1–2: Foundations
Repo scaffolding, CI/CD, core schemas, basic ingestion, DNS + WHOIS checks.

Telemetry scaffolding and error taxonomies.

2. Sprints 3–4: Connectors and ML v1

Registry connectors, LinkedIn company page verifier, text classifier MVP.

Badge issuance v1 with signed metadata.

3. Sprints 5–6: Review Console and Appeals

VerifyHire AI: Smart JOB Legitimacy System 18

 Case queue, templates, audit logs, redaction, role‑based access.

4. Sprints 7–8: Scalability and Hardening

Caching layers, circuit breakers, chaos testing, perf budgets.

5. Sprint 9+: Compliance and Enterprise Features

SSO/SAML, exportable audit, fine‑grained scopes, data retention controls.

Comprehensive Competitive Analysis

1. Categories
General job boards, ATS/HRIS vendors, identity verification providers, fraud
detection platforms.

2. Gaps Addressed by verifyHire

Cross‑platform evidence graph, signed assertions, and public badge interoperability.

3. Risks
Platform incumbents replicating core checks; mitigation via partnerships, patents,
and data moat.

Detailed Financials
1. 36‑Month P&L Sketch
Revenue: Subscriptions, usage fees, enterprise add‑ons.

COGS: Data, compute, storage, human review.

Opex: R&D, GTM, G&A, compliance.

EBITDA trajectory and cash runway under base and bear cases.

2. Cohort Economics
Logos by segment, ARPA, expansion, churn, NRR targets.

Payback periods by channel: direct vs partner‑led.

VerifyHire AI: Smart JOB Legitimacy System 19

 Risk Register (Operational)
Source dependency: Registrar API outages → multi‑source + backfill.

Abuse of badge assets → signed URLs and referrer validation.

Reviewer bias → calibration sessions and bias metrics.

Appendices
A. Sample JSON Assertion

{
"iss": "https://api.verifyhire.example",
"sub": "job:12345",
"employer": "employer:67890",
"verified_at": "2025-10-01T12:34:56Z",
"exp": 1738497600,
"checks_passed": ["dns_domain_match", "registry_match", "contact_channel_corp
orate"],
"confidence": 0.94,
"evidence_hash": "sha256-..."
}

B. Sample Webhook Payload

{
"event": "job.verified",
"id": "evt_abc123",
"job_id": "12345",
"status": "verified",
"confidence": 0.92,
"reasons": ["domain_alignment", "registry_record"],
"occurred_at": "2025-10-01T12:35:01Z"
}

C. Reviewer Playbook Excerpts

If domain misaligned, instruct DNS challenge or email from CFO/HR on corporate
domain.

If salary implausible, request internal posting URL or budget doc excerpt.

VerifyHire AI: Smart JOB Legitimacy System 20

 D. Data Retention Matrix
Evidence: 12 months default, 36 months enterprise.

Logs: 30 days hot, 365 days cold for audit trail with legal holds.

E. Metrics Definitions
Scam exposure rate: Complaints per 10k views pre/post.

Verification coverage: Verified postings / total postings.

Appeal overturn rate: Appeals reversed / total appeals.

VerifyHire AI: Smart JOB Legitimacy System 21