Own Your Staffing Data Outside Bullhorn.

Your Bullhorn data in a private AWS sandbox that you own, with a clean schema, a ready-made API and sync in under 5 minutes. About half what Bullhorn charges for Data Replication, and your data never sits on our servers.

Powering high-security workforces

Bullhorn ATSAmazon Web ServicesDatabricksAWS Bedrock
Your AWS account Your KMS keys < 5 min sync
BullhornREST API plus the fields it doesn't expose
Your AWS accountprivate VPC
Sync Lambdasrate-limit safe
PII maskingSSNs tokenized
Your private vaultAurora or Databricks · normalized schema · encrypted with your keys
REST APISwagger docs
SQL viewsBH_Candidate…
AI add-onPhase 2
Cloudgate deploy role: deploy and monitor only. No data access. Revoke it in one click.

The problem

Bullhorn charges you a fortune to get your own data back.

Bullhorn Data Replication runs $30K to $60K+ a year, and you still have to run your own SQL Server to receive it. Writing your own scripts against the REST API is worse: the API has holes, so they quietly miss data.

Self-hosted replication
  • $30K to $60K+ a year for Bullhorn's Data Replication add-on
  • Windows Server and SQL Server licenses you have to buy, patch and babysit
  • Unmasked SSNs and candidate PII copied into another database
  • A raw replica nobody can actually search or put AI on safely
Cloudgate data vault
  • About half the cost of Bullhorn Data Replication
  • A private vault in your own AWS or Databricks account, managed for you
  • SSNs and sensitive fields masked or tokenized before anything else touches them
  • Continuous, rate-limit-safe sync that keeps your live Bullhorn fast
  • A clean schema and documented API, with your existing reports still working

Platform

A private sandbox, a clean API and nothing to babysit

We replace Bullhorn Data Replication with a pipeline that runs inside your own AWS account. No Windows servers, no SQL Server licenses and no DBA on call. We deploy it, monitor it and keep it current.

True data ownership

Candidates, jobs, submissions, placements, notes and contacts sync continuously into your own private AWS or Databricks environment.

Clean schema & Swagger API

We fill the holes in Bullhorn's REST API, including fields it doesn't expose, and normalize your history into a clean schema with a documented REST API inside your VPC.

Automated PII & SSN protection

Sensitive candidate fields are stripped, masked or tokenized before analytics or any AI model can read them.

Your reports keep working

Compatibility views mirror Bullhorn's replication tables, like BH_Candidate and BH_Placement, so Power BI and Tableau just repoint. Nothing gets rebuilt.

Your data, your way

How your team actually uses it

Once your vault is live, every Bullhorn record is clean, normalized and current. Your developers and BI team use it from day one. The recruiter portal and AI assistant are optional add-ons for later.

Your vault · REST API
RequestGET /v1/candidates?near=33602&radius=25&skill=PE
// 200 OK · scoped to your API key · SSNs masked
{
  "count": 38,
  "results": [
    { "id": 90311, "name": "Candidate 1", "ssn": "***-**-4821",
      "distance_mi": 4.2, "last_submission": "2025-08-14" },
    …
  ]
}
Swagger / OpenAPIAPI keys per appInside your VPCJSON
-- Placements this year by recruiter and client
SELECT owner_name, client_name,
       COUNT(*) AS placements,
       SUM(spread_per_hour) AS weekly_spread
FROM placements
WHERE start_date >= '2026-01-01'
GROUP BY owner_name, client_name
ORDER BY placements DESC;
RecruiterClientPlacementsSpread
J. RiveraClient A14$4,210
M. ChenClient B11$3,380
A. PatelClient C9$2,940
Power BITableauExcelDatabricks SQLPython
Which licensed structural engineers within 25 miles of Tampa interviewed with us in the last two years but were never placed?
I found 38 candidates who match. The strongest five:
  • 3 interviewed with Client B in 2025 and are marked available
  • 2 hold active Florida PE licenses and updated resumes this quarter
Sources: candidates, submissions and notes tables · SSNs masked · scoped to your desk
Private model on AWS BedrockYour data never trains a modelRole-scoped answers

Optional add-on

A private AI assistant on your own data

Add an LLM layer to your vault and your team can ask questions instead of building reports. It runs inside your own cloud environment, so nothing leaves your control.

Discuss it on your audit
  • Plain-English questions. "Who did we place at this client last year?" or "Which candidates went quiet after a second interview?"
  • Private by design. Hosted on AWS Bedrock in your environment. Your data is never used to train a model.
  • Permission-aware. Every answer is limited to the records the person asking is allowed to see.
  • Masked before it reads. SSNs and sensitive fields are tokenized before the AI ever touches them.
  • Isolated memory. Each role's agent keeps its own context, so nothing bleeds between desks.

How it works

A zero-downtime cutover in about three weeks

Architecture audit

We map your Bullhorn footprint, data volumes, compliance needs and where your team loses time today.

Deployed in your account

You launch one template in your AWS account. It creates the private vault, masking rules, your encryption keys and our limited deploy role.

Run side by side

Weeks 1 and 2: your full history backfills without choking your API quota while your current replication keeps running. We match row counts table by table.

Repoint and cancel

Week 3: your dashboards point at the new views, live sync runs under 5 minutes, and you drop Bullhorn Data Replication at renewal.

Security

Enterprise AI without a compliance breach

Your data never sits on our servers. Cloudgate deploys into your own cloud account, and every person and agent only sees what their role allows.

  • Your account, your keysWe deploy into your own AWS or Databricks account. You hold the root credentials and can revoke our access in one click. No shared databases and no cross-tenant access.
  • Masked before it landsSSNs and sensitive identifiers are tokenized in the pipeline, so analytics and AI never see raw values.
  • Role-based accessRecruiters, managers and AI agents query only the desks and records they're authorized for.
  • Isolated AI memoryAWS Bedrock agents run with isolated memory per role, so context never bleeds between teams.
  • Encrypted with keys you ownEncrypted in transit and at rest with customer-managed AWS KMS keys. Our support access is time-bound and logged.
  • A firewall in writingData ownership, non-use and non-solicit terms in every agreement, covering Cloudgate and Altus alike. Your data is yours, period.

See the full security architecture

Case study · Altus Jobs

Built inside a staffing firm, not a lab

We were paying for our own data twice. Once to store it in Bullhorn and again to get it back out. So we built the way out.

Altus Jobs, an Orlando recruiting firm specializing in architecture, engineering and the built environment, has run its Bullhorn data on the Cloudgate engine for years. It worked well enough that Altus is now sunsetting Bullhorn entirely and saving $48,000 a year at 30 seats. Its full history lives in its own cloud environment, pipeline activity syncs continuously, and recruiters search candidates by radius and ask questions of their own data in plain English.

$48K/yrsaved at 30 seats by sunsetting Bullhorn
~50%less than Bullhorn's Data Replication fee
< 5 minfrom a Bullhorn update to your private vault
6+ yrsin daily production on a live Bullhorn account

Leadership

Operators who live in Bullhorn every day

Saum Sharifi

Saum Sharifi

Co-Founder & CEO

Saum leads Cloudgate's product vision, enterprise strategy and staffing industry adoption. For more than 15 years he has run Altus, a nationwide executive search firm for architecture, engineering and construction, and he saw firsthand how ATS reporting limits, fragile replication setups and rigid keyword search make agencies pay again for talent already in their own database. Cloudgate was built to fix that inside Altus first.

  • Founder and CEO of Altus Jobs for 15 years, operating on Bullhorn every day
  • Co-founder of Cloudgate Technologies
  • Partner at PKG, a holdings company that acquires and manages restaurant franchises
  • Early seed investor and advisor to Advocacy, a litigation AI platform
  • Built Cloudgate inside Altus until Bullhorn became optional: Altus is sunsetting it and saving $48,000 a year
StrategyPartnershipsClientsOperations
Chris Callaway

Chris Callaway

Co-Founder & CTO

Chris has spent more than 15 years as a software engineer connecting staffing technology to the systems firms already run. He architected the Cloudgate sync engine that has kept Altus Jobs' Bullhorn data in its own cloud for years, and today designs agentic AI on AWS for enterprise clients. He leads platform architecture, data security and every client deployment.

  • Early engineer at Fyre, where he helped build VMSync, the VMS-to-ATS sync product (acquired in 2016)
  • Co-founder and lead architect of Cloudgate: recruiter automation, real-time pipeline boards and Bullhorn data mirroring
  • About 9 years of enterprise Bullhorn integrations: 100+ integrations across platforms like Workday, HubSpot and Workato, including systems with no public API
  • Has built production systems on heavily regulated data: HIPAA-compliant hospital records, proprietary stock market aggregates and payroll with Social Security numbers and tax records
  • Builds Amazon Bedrock and AgentCore AI agents for healthcare and financial clients, with guardrails and per-client data isolation
  • Runs his own multi-agent AI operations platform built on Claude
BullhornAWSSecurityAI agents

Architecture audit

See what your Bullhorn data could do outside Bullhorn

A 30-minute session with our team. No pitch deck, just your stack.

  • Map your current replication or export setup
  • Identify PII exposure and compliance gaps
  • Estimate what's buried in your candidate database
  • Leave with a target architecture, whether or not you work with us

We reply within one business day. Your details are never shared.