Skip to content
Altaarx
§ 01 — Trust
MMXXVI

Our trust posture, in plain language.

What we promise. What we sign. What we do. What we do not yet have. This page is meant to be readable by a counterparty's general counsel, a compliance officer preparing for examination, and an AI auditor looking for evidence — without gates, NDAs, or “request access” forms.

Last reviewed: April 28, 2026. Reviewed quarterly. Change log at the bottom of this page.

§ 02 — Commitments

Four things we commit to in writing.

The shortest possible answer to the most common vendor due-diligence questions. Each of these is a contractual commitment in our standard engagement, not a marketing line.

  1. ICommitment

    Business Associate Agreement on request, signed before kickoff

    Any engagement involving Protected Health Information includes a HIPAA Business Associate Agreement signed before any access to PHI. We use the Common Paper BAA template, customized for AI-specific data flows. Subcontractor BAAs are executed with any sub-processors that may handle PHI, with the same restrictions and conditions that apply to Altaarx. We routinely engage with mid-market medical practices and digital-health startups under HIPAA scope.

  2. IICommitment

    72-hour incident notification under Reg S-P

    When a covered institution — registered investment adviser, broker-dealer, investment company, transfer agent — engages Altaarx, Altaarx is a "service provider" under SEC Regulation S-P §248.30(b)(4). Our engagement contract binds us to written notification of any unauthorized access to or use of Customer Information within 72 hours of discovery, in the form and content the covered institution requires for its own 30-day customer-notification obligation. This commitment is contractual, not aspirational.

  3. IIICommitment

    Mutual NDA available before any sensitive disclosure

    A mutual non-disclosure agreement is signed before any discussion that requires disclosure of confidential information — system architecture, data flows, prior incidents, board minutes, regulatory correspondence. Our standard MNDA is reciprocal, three-year survival on confidentiality obligations, with carve-outs for residuals, independently developed information, and information rightfully obtained from third parties. Available on request.

  4. IVCommitment

    Client data stays in the client's environment by default

    Our default engagement model is to operate inside the client's tenant — Google Workspace, Microsoft 365, AWS, GCP, Azure, on-premises — under a least-privilege account provisioned by the client and revoked by the client at engagement end. We do not require copies of client data on Altaarx infrastructure. Where a working copy is genuinely necessary (rare; documented case-by-case), it is encrypted in transit and at rest, on hardware-key-protected, FileVault-encrypted endpoints, and is destroyed or returned at engagement end with written certification.

§ 03 — Frameworks

The standards every engagement maps to.

Each engagement is mapped to whichever subset of these frameworks actually examines the client. We don't promise certifications we don't hold; we deliver work product that survives examination under the frameworks below.

  • NIST Cybersecurity Framework

    v2.0 (Feb 2024)

    The baseline cyber-risk vocabulary every regulated buyer references. Our audit deliverable maps controls to the six functions: Govern, Identify, Protect, Detect, Respond, Recover.

  • NIST AI Risk Management Framework

    AI 100-1 + Generative AI Profile AI 600-1

    The U.S. de facto standard for AI governance. Required mapping for any engagement involving generative or agentic AI.

  • ISO/IEC 27001:2022

    Information security management

    Used when the client carries — or is preparing for — ISO 27001 certification, especially European subsidiaries and multi-jurisdictional family offices.

  • ISO/IEC 42001:2023

    AI management systems

    The first international standard for AI management systems. Increasingly cited in vendor contracts and procurement frameworks.

  • HIPAA Security Rule

    45 CFR §§ 164.302–164.318

    Administrative, physical, and technical safeguards for ePHI. The 2026 NPRM updates remove "addressability" — MFA, encryption, and segmentation become required, not optional.

  • PCI DSS

    v4.0.1 (current)

    Required for any payment-card data path. Our work explicitly avoids putting AI in the cardholder data environment unless requirement-by-requirement mapping is documented.

  • GDPR

    Regulation (EU) 2016/679

    Applied when the client serves EU data subjects, including U.S.-based family offices with European LP investors or beneficiaries.

  • SEC Regulation S-P

    17 CFR Part 248, as amended May 16, 2024

    Smaller adviser compliance June 3, 2026. Service-provider oversight, 72-hour breach notification, 30-day customer notification.

  • CIS Controls

    v8.1

    Practical implementation of the higher-level frameworks. Used as a check on our own audit deliverable for completeness.

  • OWASP Top 10

    2021 (web) + LLM Top 10 v2 (2025)

    Web-application security and LLM-specific risks: prompt injection, insecure output handling, training-data poisoning, supply-chain attacks, sensitive-information disclosure, model denial-of-service.

  • U.S. Treasury FS AI RMF

    Released February 19, 2026

    Financial Services AI RMF aligned with SR 11-7 model risk management. Newly relevant for RIA, broker-dealer, and bank-affiliated client engagements.

§ 04 — Reg S-P

For RIAs, broker-dealers, and covered institutions.

If Altaarx is engaged by a covered institution under SEC Regulation S-P, Altaarx is a “service provider” under §248.30(b)(4). Below is what your written service-provider oversight policy gets when you engage us. This section is written for the compliance officer preparing the next exam and the general counsel reviewing the MSA.

What we sign:

  • Written service-provider agreement that meets the Amendments' content requirements: scope of access to Customer Information, security obligations, breach notification (72 hours, written), customer-notification responsibilities, return/destruction of Customer Information at engagement end.
  • 72-hour written notification of any unauthorized access to or use of Customer Information, with sufficient detail to support the covered institution's own 30-day customer-notification analysis under §248.201(b).
  • Cooperation obligations for the covered institution's investigation, evidence preservation, and the covered institution's recordkeeping requirements under §248.30(b)(5).
  • Subcontractor flow-down: any sub-processor that touches Customer Information is bound to equivalent or stronger commitments before access is granted. Sub-processor list is on this page.
  • Recordkeeping support: Altaarx maintains internal records of access, incidents, and disposal events, and produces them on the covered institution's request to support the institution's own five-year retention obligation under the Amendments.

What we don't sign: unilateral indemnification clauses untethered from cause; obligations beyond our service scope; provisions that would require us to violate other regulatory obligations (e.g. preserving evidence required by another regulator).

§ 05 — HIPAA

For covered entities and business associates.

Altaarx is a HIPAA business associate when engaged by a covered entity or by another business associate to perform functions involving PHI.

Standard BAA terms. Our standard BAA mirrors the HHS sample BAA at 45 CFR § 164.504(e), with AI-specific additions: explicit prohibition on using PHI to train any model not exclusively governed by the covered entity's instructions; explicit handling of inference-time data residency for any AI vendor used in the engagement; and explicit subcontractor flow-down language to any sub-processor that may incidentally access PHI.

Subcontractor BAAs. Altaarx executes downstream BAAs with any sub-processor that creates, receives, maintains, or transmits PHI on Altaarx's behalf, before access is granted. We will not subcontract to a sub-processor that refuses to sign a HIPAA-compliant BAA.

AI-vendor BAAs. When an engagement requires use of a major AI vendor with PHI access, we work only with vendors that offer signed BAAs covering the relevant services. As of this writing, that includes Anthropic (covered enterprise products), Microsoft Azure OpenAI Service, Google Cloud Vertex AI, AWS Bedrock for HIPAA-eligible models, and OpenAI for ChatGPT for Healthcare and API healthcare customers. Vendor BAA scope is verified per engagement; we don't assume a BAA covers a service it doesn't.

2026 NPRM-readiness. The HHS Notice of Proposed Rulemaking (Jan 6, 2025; final rule expected mid-2026) introduces mandatory MFA, mandatory encryption, mandatory network segmentation, annual technical testing, 24-hour incident reporting from BAs to covered entities, and annual BA verification of technical safeguards by a subject-matter expert. Our engagement deliverables explicitly address each of these where they apply to the AI surfaces in scope.

§ 06 — Operational security

What our day-to-day looks like.

The mechanics. None of this is speculative — these are the controls in place on the engagement infrastructure as of the date at the bottom of this page.

Default engagement environment
Work inside the client's tenant. Least-privilege account provisioned by the client; revoked by the client at engagement end. Altaarx does not require copies of client data on Altaarx-controlled infrastructure.
Endpoint security
macOS endpoints with FileVault full-disk encryption, Gatekeeper code-signing enforcement, XProtect malware scanning, automatic OS security updates, and a dedicated client-engagement profile separate from personal use. iOS / iPadOS endpoints with passcode + biometric, Find My, and remote wipe.
Authentication
Hardware-key-enforced MFA (FIDO2 / WebAuthn) on every account that touches client systems or sensitive infrastructure. Phishing-resistant by design. No SMS-based 2FA for any privileged account.
Credential storage
1Password, with dedicated vaults per client engagement. Master credentials and recovery keys stored in secure offline backup. No credentials in chat, email, code, or unencrypted notes.
Encryption in transit
TLS 1.2+ on every connection. TLS 1.3 preferred where supported. SSH over the strongest mutually-supported cipher (Ed25519 / Curve25519 / AES-256-GCM). Strict DNS resolvers (DNS-over-HTTPS or DNS-over-TLS where infrastructure permits).
Encryption at rest
FileVault on macOS endpoints. Encrypted git repositories (git-crypt or equivalent) for any work product that contains sensitive data. Cloud-side encryption-at-rest for all storage backends in scope (AES-256).
Network
No client-data transit over consumer Wi-Fi without VPN. Travel posture: no work on hotel or airport networks; mobile hotspot with WPA3 only.
Backup
Time Machine to a hardware-encrypted volume, plus encrypted off-site backup of work-product repositories. Backup encryption keys held separately from backup media.
Logging
Engagement work performed within client systems is logged by the client's existing audit infrastructure. Where Altaarx maintains its own working copy (rare), access is logged locally and the log is provided on request.
Engagement scoping document
Every engagement begins with a written scoping document specifying: data categories in scope, systems in scope, access pattern (federated, copy, observe-only), retention period for any working copies, destruction or return procedure at engagement end, breach-notification path, and SLA on response. Signed by both sides before access.
Disposal
Working copies of client data are destroyed at engagement end via cryptographic erasure (key destruction) for cloud-side storage and secure-delete for local working copies. Written certification of destruction provided on request.
§ 07 — Sub-processors

The vendors behind Altaarx itself.

Altaarx is small but not solo at the infrastructure layer. Below is the current list of vendors and tools that support Altaarx's own operations. This list is updated when sub-processors are added, changed, or removed. Material changes are noted in the change log at the bottom of this page.

  • Cloudflare, Inc.
    DNS, CDN, hosting (Workers), bot protection (Turnstile), Web Analytics
    US (San Francisco, CA)
    Yes — Cloudflare DPA + BAA on enterprise plans where required
  • Resend
    Transactional email (contact form notifications)
    US (Delaware)
    DPA available
  • GitHub, Inc.
    Source code repositories (private), engagement project boards
    US (San Francisco, CA) — Microsoft subsidiary
    Microsoft DPA + BAA available where required
  • Atlassian
    Project tracking (Jira), internal documentation (Confluence)
    Australia / US (San Francisco, CA)
    Atlassian DPA
  • 1Password
    Credential storage, dedicated vaults per engagement
    Canada (Toronto)
    DPA available
  • Apple, Inc.
    macOS / iOS endpoints, iCloud Drive (for Altaarx-internal documents only — not client data)
    US (Cupertino, CA)
    Standard Apple business terms
  • Anthropic, PBC
    Internal use of Claude for Altaarx's own work product (drafting, research). Not used to process client data unless explicitly authorized in the engagement scoping document.
    US (San Francisco, CA)
    Anthropic DPA + BAA on commercial enterprise products

Vendors used only for Altaarx's own administrative operations (banking, accounting, telecom) and that do not touch client data are intentionally not listed. Engagement-specific sub-processors — for example, an AI vendor required by a particular client deployment — are documented in the engagement scoping document and bound by engagement-specific BAA / DPA terms before access.

§ 08 — Insurance + contracts

What's behind the work.

Standard contracts and insurance coverage provide the legal and financial backstop a sophisticated counterparty expects from any service provider.

Insurance. Errors and Omissions (Professional Liability), Cyber Liability, and Commercial General Liability coverage is bound through a Florida-licensed broker. Coverage limits, named insured, and certificate of insurance available on request to engaged or prospective clients under MNDA.

Standard contracts. Master Services Agreement, Statement of Work, Mutual NDA, Data Processing Addendum, and Business Associate Agreement templates are derived from Common Paper standards (commonpaper.com) — vetted by general-counsel-grade reviewers and widely recognized in venture-backed technology procurement — and customized for AI-specific data flows and AI-specific governance language. The master MSA has been reviewed by Florida counsel.

Counterparty entity. Engagement contracts are signed in the name of the parent Delaware C corporation, foreign-qualified in Florida, doing business as Altaarx. Certificate of incorporation, certificate of existence, and Florida foreign-qualification certificate available on request.

§ 09 — Vulnerability reporting

Coordinated disclosure.

Altaarx welcomes security research conducted in good faith on altaarx.com and any Altaarx-operated infrastructure. We follow a standard 90-day coordinated disclosure window.

How to report. The canonical machine-readable disclosure path is at /.well-known/security.txt, per RFC 9116. Human-readable: contact via the homepage form, with Subject: security in the message field, or use the contact path published in security.txt.

Scope. altaarx.com, mail-sending domains under altaarx.com, and any Altaarx-operated infrastructure clearly attributable to Altaarx. Out of scope: third-party services (Cloudflare, Resend, etc. — please report to those vendors directly), social-engineering reports, and physical security of any individual.

What we ask. Provide reproducible proof-of-concept; do not access, modify, or delete data that does not belong to you; do not degrade service availability; give us 90 days to remediate before public disclosure.

What we commit. Acknowledgment of the report within 5 business days; a substantive response within 15 business days; coordinated public disclosure once a fix has shipped or after 90 days, whichever comes first; public credit on this page if the reporter requests it.

§ 10 — Year-1 transparency

What's not yet here, and why.

Altaarx is a new entity. Some maturity signals take time. Listing them honestly is itself a trust signal — the firm you should be most worried about is the one claiming a posture it hasn't earned yet. Below is what we don't have, why, and when.

SOC 2 Type II
Planned: Year 2. Type I scoped first (~6 months from launch), Type II (12-month observation period) following. Initial audit budget ~$30,000–$50,000.
ISO/IEC 27001 certification
Planned: Year 3, if a client engagement materially benefits from Altaarx-side certification (most engagements do not, because controls live in the client's environment).
ISO/IEC 42001 certification
Considering for Year 3. The standard is new (2023); auditor capacity is still maturing; the value to small consultancies is debated. We will revisit annually.
Penetration test report
Planned: Year 1, post-first-engagement. The site itself has been audited by a single qualified practitioner (the founder, with PenTest+); a third-party pen-test will be commissioned once there is engagement-controlled infrastructure that justifies the spend.
Public client roster
None yet — Altaarx launched April 2026. First public case studies expected Q4 2026 once first engagements complete and clients give written consent for attribution. We will not publish client names without written consent, ever.
Bug bounty program
Considering after first engagement. Coordinated disclosure (above) is in place now; a paid bounty program is overhead that does not yet match the attack surface.
24/7 staffed SOC
Out of scope for Altaarx's service model. Altaarx is not a managed-security provider; clients with 24/7 monitoring requirements should engage an MSSP in addition to Altaarx.
Multi-region data residency
Not applicable — Altaarx does not by default store client data on Altaarx-controlled infrastructure. Where a working copy is required, residency is documented in the engagement scoping document.

An examiner cannot audit a vendor that hides its limitations. We would rather lose an engagement to a buyer who needed something we don't have than win one we can't deliver against.

§ 11 — Verification

Where to check.

Independent verification paths for the claims on this page.

Founder identity and credentials
ORCID iD 0009-0006-6888-5396 (registered 2026-04-28). Capitol Technology University doctoral registration verifiable through the institution's registrar. WGU degrees verifiable through the National Student Clearinghouse. CompTIA certifications (Sec+, CySA+, PenTest+) verifiable at verify.comptia.org. ISC2 SSCP verifiable at verify.isc2.org.
Entity
Delaware Division of Corporations entity search (icis.corp.delaware.gov). Florida foreign qualification verifiable at search.sunbiz.orgonce filed. Florida fictitious name “Altaarx” verifiable at the same URL once registered.
Domain control
DNS records at altaarx.com publicly resolvable. SPF, DKIM, DMARC records published. SSL certificate issued via Cloudflare and verifiable at crt.sh.
Documents on request
Certificate of insurance, BAA template, MSA / SOW templates, MNDA template, sub-processor list (current version), engagement scoping document template — provided to engaged or prospective clients on request, under MNDA where the document contains commercial terms.
Direct line
A real person responds to inquiries via the contact form on the homepage. Replies within one business day. Founder is the responder for any pre-engagement question.
§ 12 — Change log

Page version and history.

This version: v1.0 — Published April 28, 2026.

Reviewed quarterly minimum. Material changes — adding or removing a sub-processor, changing a binding commitment, updating a regulatory effective date — are logged below with the date of change. Editorial changes (typo fixes, link refreshes) are not logged.

  1. v1.0 published. Initial trust posture document covering frameworks, Reg S-P and HIPAA service-provider commitments, operational security controls, sub-processor list, insurance and contracts posture, vulnerability reporting, year-1 transparency disclosure, and verification paths.