Roger Neal

• Evaluate and validate large language model (LLM) outputs across different domains.

Senior Cloud Security Consultant (Jul 2022 to Jan 2026); Cloud Security Consultant (Oct 2020 to Jul 2022)

• Served as a cloud security SME to enterprise-level clients across AWS, Azure, and GCP. Led cloud security assessments and Cloud Controls Matrix (CCM) engagements to identify risks and deliver prioritized, risk-based remediation aligned to 20+ compliance frameworks.
• Designed and implemented secure cloud architectures using Terraform and AWS CloudFormation, enabling scalable, repeatable deployments and accelerating customer adoption.
• Led threat modeling initiatives using the PASTA methodology, translating complex technical risks into actionable insights for engineering teams and executive stakeholders.
• Produced and delivered video and audio technical content as a cloud security expert.
• Acted as a trusted advisor to enterprise clients, delivering technical demonstrations and go-to-market content that translated security capabilities into measurable business value.
• Automated security controls and integrations using Python.

• Executed a full testing lifecycle, including functional, regression, system, performance, and UAT testing within an agile delivery team. Developed and maintained test scripts in IBM Rational Clear Quest (RQM) and CA Service Desk, tracking and reporting defects, and preserving testing artifacts in SharePoint for auditing and compliance.
• Served as client services liaison between the Veterans Administration (VA) and mortgage services, ensuring accuracy and integrity of VA reporting deliverables.
• Provided technical business consulting, formulating systems scope and objectives aligned to client business plans and bridging communication between end-users and developers on the implications of technology solutions.

Native Rust statistical computing library implementing 24 regression, survival analysis, and time-series modules.

• Delivered full model summaries (SEs, CIs, p-values, AIC/BIC) plus a formula API (y ~ x1 + C(group)).
• Docker-benchmarked 24 estimators against Python statsmodels (10k rows, matched arm64 containers): faster on 16/24, from 2× (logistic/probit, IRLS-bound) up to 148× (Cox PH, where lifelines' general-purpose optimizer carries heavy Python overhead).

Contributed 3 merged pull requests to Hummingbird, a production Swift server-side framework with 1,790+ GitHub stars, spanning test infrastructure, Codable extensions, and middleware features.

• PR #869: Designed and shipped an opt-in serveOnNotFoundResponse flag for Hummingbird's FileMiddleware, closing a gap where routes returning a 404 Response (vs. throwing) skipped static-file fallback; preserved default behavior and added GET/HEAD/edge-case test coverage.
• PR #870: Added FormatStyle/ParseStrategy-based date encoding/decoding to Hummingbird's URLEncodedFormEncoder/Decoder, providing a thread-safe, Sendable alternative to DateFormatter while maintaining full backward compatibility; included round-trip unit tests.
• PR #877: Hardened Hummingbird's HTTP/2 test suite by adding explicit TLS ALPN protocol negotiation and protocol-version assertions, closing a coverage gap where tests could silently pass without actually running over HTTP/2.

• Participated in workgroup discussions and meetings focused on advancing Swift as a supported platform for Android development.