Matt Grofsky.

AI Safety
& Security.

Technology and security leader with an operator’s mindset and an architect’s curiosity, comfortable moving from executive strategy into system design, threat modeling, AI security, and technical validation.

Explore the workGitHub

Selected work.

Research, platforms & applied AI

Implementation approach

Human-led architecture, planning, security review, and QA; LLM-assisted implementation.

Technical practice.

Engineering & applied research

Hands-on security engineering & research

Designs and validates AI-assisted security systems, personally operates vulnerability disclosure programs, and works directly on incident investigation and remediation verification.

Architecture & review

Defines system architecture, permission boundaries, acceptance criteria, and tests. Guides AI-assisted implementation through security review, code review, and runtime checks.

Investigation & evaluation

Validates vulnerability reports, investigates incidents, and checks remediation. Applies cybersecurity analytics and evaluates model outputs against reference material and explicit criteria.

Experience

Jan 2026 – Present

Sharpen Technologies

CISO

Leads security across Sharpen, Plum Voice, Webtext, and Ytel. Runs and manages R&D and IT across the four businesses, leading four managers and a 22-person organization.

Role detail

Chief Information Security Officer

Leads security across Sharpen, Plum Voice, Webtext, and Ytel. Runs and manages R&D and IT across the four businesses, with responsibility for engineering, CloudOps, and product management. Following Sharpen's acquisition of Ytel, established a shared security operating model and maturity rubric across the four businesses, separating documented controls from operating effectiveness and turning findings into prioritized engineering and integration roadmaps.

Runs the security function as its sole dedicated practitioner under PCI DSS, HIPAA, and SOC 2 obligations. Leads secure design review, SAST, SCA, secrets and container scanning, external penetration testing, SIEM and endpoint security, incident response, risk acceptance, and verified remediation.

Uses concurrent traffic-split tests to evaluate reliability changes. Requires clear metric definitions, sources, time windows, and missing-data handling before acting on operating data.

Reports to the CEO and manages four managers and a 22-person organization across Sharpen, Ytel, Plum Voice, and Webtext.

Product management sets delivery priorities. Matt makes the final call when priorities conflict, requiring options to clear risk and reuse standards before weighing cost-benefit return, impact against complexity, and mission alignment.

Coaches engineers to own SLOs, SLIs, and error budgets. Across the career: roughly 200 hires over 20 years, with organizations of up to 40 people.

Jul 2025 – Present

The George Washington University

Professorial Lecturer & Doctoral Advisor

Doctoral research advising in cybersecurity analytics, AI, and automation.

Role detail

Professorial Lecturer in Engineering and Applied Science; Doctoral Advisor, Cybersecurity Analytics

Advises Doctor of Engineering candidates conducting applied cybersecurity research at the intersection of analytics, AI, and automation. Guides problem framing, methodology, evaluation, and practical application, and serves as a voting member on praxis defense committees.

Jan 2011 – Jan 2026

Ytel, Inc.

CTO

Led the entire technology organization, including engineering, infrastructure, operations, and security, for CPaaS and CCaaS platforms serving more than 5,000 customers.

Role detail

Chief Technology Officer

Led the entire technology organization, including software engineering, platform architecture, infrastructure, operations, and security. Built and led multidisciplinary teams supporting CPaaS and CCaaS platforms serving more than 5,000 customers. Established the platform's security and trust foundation across cloud architecture, incident response, STIR/SHAKEN, TCPA, and protection of PII and CPNI.

Jan 2011 – Present

F9 Group, Inc.

Cofounder & Technology and Security Advisor

Secure architecture, security programs, responsible AI, and product strategy.

Role detail

Cofounder & Technology and Security Advisor

Advises SaaS, contact-center, and cloud organizations on secure architecture, security programs, incident preparedness, responsible AI, and product strategy through a two-officer consulting company.

2001

NQL

Professional Services Developer

Customer implementations for intelligent agents and content aggregation.

Role detail

Professional Services Developer

Delivered customer implementations using Network Query Language, an early purpose-built language for developing intelligent agents and content aggregation applications.

More work.

Security operations, platforms & analytics

Agentic AI security

Pentest Arsenal: Agent Controls

A private platform for authorized security testing that constrains AI-generated actions through target validation, allowlists, execution budgets, shell-free execution, audit records, and analyst review.

Technical detail

Supports vulnerability validation, API fuzzing, packet analysis, defensive verification, and ranked findings for red-team and blue-team work.

Constrains AI-generated actions through target validation, phase-specific allowlists, execution budgets, shell-free execution, audit records, and analyst review.

Incident response

Incident Investigation & Malware Analysis

Live incident response and phishing-payload analysis, including VirusTotal submissions as the initial reporter.

Technical detail

Incident response

Built an AI workspace for evidence collection, persistence hunting, containment checks, control-gap analysis, remediation tracking, and reporting. Analysts control high-impact actions and final conclusions.

Malware analysis

Used LLMs to assist with decompiling phishing payloads and submitted samples to VirusTotal, in some cases as the initial reporter.

Security operations

Vulnerability Disclosure & Remediation

Established and personally operated HackerOne vulnerability disclosure programs for Ytel and Sharpen.

Technical detail

Reviewed researcher submissions, reproduced or validated reported issues, and assessed reachability, likelihood, and business impact. Aligned priorities with engineering and created Jira remediation work.

Maintained researcher communication through disposition and engineering follow-through. Turned validated findings into AI-assisted remediation and regression tests that other engineers could run.

SOC prototype

AI-Assisted Security Alert Triage

A prototype that adds context and supporting reasoning to SIEM alerts for analyst review.

Technical detail

Suggests severity, false-positive likelihood, and analyst next steps.

Does not suppress alerts or make disposition decisions.

Compliance automation

PCI, HIPAA & SOC 2 Mock Auditor

A read-only AI assessor that separates documented policy from operating evidence.

Technical detail

Examines live cloud environments, infrastructure as code, and application repositories against PCI DSS, HIPAA, and SOC 2.

Private analytics

Security & Operational Analytics

Security and operational analytics with explicit checks on data quality, uncertainty, and model evaluation.

Technical detail

Operating data

Python, SQL, and BigQuery workflows with weighted weekday baselines, two-proportion tests, and effect-size thresholds. Separate application and network denominators; explicit flags for low volume, insufficient history, missing data, and failed sources.

Model evaluation

Designed an MLP workflow with AI-assisted implementation: feature preparation, scaling, cross-validation, and held-out evaluation. Used call duration as an engagement proxy; owned analytical design, validation, and interpretation.

Platform engineering · 2016 public coverage

Ytel CPaaS (formerly message360)

Led development of a communications platform built around Ytel's APIs. Guided developers and partners from architecture through production, turning feedback into documentation, examples, and reusable integration patterns.

Read the 2016 platform interview
Technical detail

Custom integrations were consuming engineering capacity. A shared API platform gave customers more room to build and the engineering team more room to work on capabilities shared across customers.

Across Ytel's developer platform, guided developers, customers, and partners from architecture discovery through API integration and production rollout. Feedback informed documentation, reusable examples, platform capabilities, and roadmap priorities.

Ytel's wider developer platform spanned voice, messaging, email, customer integrations, and developer tooling, supported by documentation, code examples, and SDK patterns across 19 programming languages.

Product engineering · 2021 public coverage

Ytel Workflow

Led work on a visual workflow builder for communications.

Read the 2021 product coverage
Technical detail

Enabled customers to configure call routing, messaging, and conversation tracking without custom code for each flow.

Combined visual orchestration, webhooks, routing, and communications APIs.

Applied AI · 2019–2020 work

Earlier Automation & Applied AI

Automated alerts in 2019, followed by computer vision, call analysis, and speaker diarization in 2020.

Read the 2020 technical article
Technical detail

Call analysis combined speech recognition, speaker diarization, keyword detection, and automated alerts to surface compliance and scam indicators. Worked with engineering and operations teams to make call data useful for investigation and service improvement.

Research, writing & coverage.

Google Scholar

Research

Technical writing

Interviews & coverage

MediumORCID

Education & credentials

  • D.Eng., Cybersecurity Analytics — The George Washington University
  • M.Eng., Cybersecurity Policy & Compliance — The George Washington University
  • B.A., East Asian Cultures — University of California, Irvine
  • CISSP — Certified Information Systems Security Professional
  • Co-inventor, U.S. Patent 7,831,028, System and Method for Call Tracking (granted 2010).

Professional community

  • Member, InfraGard
  • Member, CR-ISAC
  • Member, IEEE Computer Society
  • Member, CyberRisk Alliance Expert Community
  • Advisory Board, Chatsworth High School G+STEAM Magnet (LAUSD), 2019–Present

Security frameworks

Applied MITRE ATT&CK mapping in AegisShield’s threat-model evaluation; familiar with D3FEND and ATLAS.

Research & intelligence resources

Cyber Threat NewsVendor IntelligenceGovernment SourcesResearch ToolsAI/ML Threat IntelligenceTemporal RAG Research Code

Continue the conversation.

LinkedInGitHub