[FIG. 01 // SUBJECT] SCALE: 1:1
Clark Ohlenbusch Portrait
REF: ARCHIVE-2026 PORTRAIT
[INDEX // 001]

AGENTIC AI SECURITY · VOICE AI · DISTRIBUTED ARCHITECTURE · BOSTON, MA

+ [SECTION // 02 · CHRONOLOGICAL TRAJECTORY]
RANGE: 2023 - PRESENT 06 CHAPTERS · PRESENT → OLDEST

 

[06 ARCHIVED CHAPTERS] SCROLL TO REVEAL TIMELINE ↓
[CH. 01 // SAVIYNT INC.] CURRENT OCCUPATION
[AUG 2023 - PRESENT · BOSTON, MA / REMOTE]

[ROLE] Product Manager / Forward Deployed Engineer
[DOMAIN] Agentic AI Security (Zuma Platform)
[ENTERPRISE IMPACT] ~$15M Closed-Won Deal Value
  • [01] Product manager and technical lead for Zuma (Saviynt's agentic AI security platform) through GA; authored Builder Experience PRFAQs for agent SDK/CLI, dev-to-prod promotion workflows, and secret scanning.
  • [02] Architected runtime authorization engine integrating OPA/Rego access control, intent-deviation detection, and behavioral monitoring; led live technical POCs to enterprise clients including Northern Trust.
  • [03] Defined enterprise Segregation of Duties (SOD) violation strategy for multi-agent chains, enforcing design-time governance and runtime boundary controls across A2A and H2A workflows.
  • [04] Ran Zuma's first live customer trial (FIS), structuring trial scope and conversion funnel; contributed technical workstreams to 22+ enterprise RFPs/RFIs, securing ~$15M in closed-won deal value.
  • [05] Led feature requirements for JWT-signed pre/post-tool webhooks, visual policy builder UX, and multi-cloud agent discovery across AWS Bedrock, Azure AI, GCP Vertex AI, Copilot Studio, and Agentforce.
  • [06] Instrumented Gateway KPIs/telemetry with Tableau reporting; delivered 'Agentic50' training to 100+ global SEs on MCP configuration, Okta to Salesforce OAuth token exchange, and Copilot Studio.
AGENTIC AI SECURITY ZUMA OPA / REGO MCP PROTOCOL AWS BEDROCK AZURE AI GCP VERTEX COPILOT STUDIO
[CH. 02 // SPOQEN] VOICE AI STARTUP
[SEP 2025 - PRESENT · BOSTON, MA / REMOTE]

[ROLE] Co-Founder / Product Lead
[PLATFORM] Voice AI Prompt Builder & Canvas
[DEPLOYMENT] ResNexus CRM Integration (Live)
  • [01] Built a voice AI prompt builder (transcript-to-flow automation, visual canvas, diff tracking/version control) with authentication and payments.
  • [02] Secured and shipped first commercial deployment: voice booking agent integrated with ResNexus CRM (VAPI + browser automation) for RV park clients.
VOICE AI VAPI VISUAL CANVAS TRANSCRIPT-TO-FLOW VERSION DIFFING RESNEXUS CRM TYPESCRIPT
[CH. 03 // MENO] LEARNMENO.COM
[2025 - 2026 · LIVE AT LEARNMENO.COM]

[ROLE] Creator & Lead Architect
[TAGLINE] "A More Human Way to Learn. The Return of Spoken Learning."
[FOUNDATION] Voice AI-Based Dialectic Education
  • [01] Architected a voice AI-based conversational education platform centered on spoken dialectic learning, natural Socratic tutoring dialogues, and real-time verbal comprehension feedback.
  • [02] Built low-latency voice pipelines enabling natural verbal interruptions, real-time Socratic inquiry, and personalized knowledge retention scoring.
VOICE-FIRST LEARNING SOCRATIC AGENTS LOW LATENCY PIPELINE CONVERSATIONAL AI EDUCATION TECH
[CH. 04 // CALICO CARE] HEALTHCARE AI CONSULTING
[OCT 2024 - PRESENT · BOSTON, MA]

 

[ROLE] Consultant (AI Development, part-time)
[DOMAIN] Clinical Voice AI & Patient Outreach
[STANDARDS] HIPAA-Compliant Architecture
  • [01] Designed HIPAA-compliant voice AI architecture (VAPI, AWS S3) to replace legacy Alexa-based COPD/CHF patient outreach protocols.
  • [02] Built automated testing suite (12+ clinical scenarios) for safety, compliance, and clinical response quality.
  • [03] Added multilingual conversational voice support (Portuguese, German) and assessed electronic medical record (EMR) integrations.
HIPAA COMPLIANCE VAPI AWS S3 CLINICAL TESTING MULTILINGUAL AI EMR INTEGRATION
[CH. 05 // TOWN OF WEYMOUTH] SECURITY CONSULTING
[MAY 2025 - AUG 2025 · WEYMOUTH, MA]

[ROLE] Security Engineering Consultant (Freelance)
[SYSTEMS] Active Directory + ADFS
[HARDWARE] Enterprise YubiKey MFA Rollout
  • [01] Designed and deployed a hardware YubiKey-based MFA solution across a Windows enterprise environment (Active Directory + ADFS).
  • [02] Configured and validated end-to-end authentication workflows; supported municipal IT staff through rollout, troubleshooting, and user enablement.
  • [03] Created internal documentation and operational runbooks to support ongoing identity administration and incident response.
YUBIKEY MFA ACTIVE DIRECTORY ADFS AUTHENTICATION WORKFLOWS ENTERPRISE SECURITY
[CH. 06 // UMASS BOSTON] TEACHING & RESEARCH
[JAN 2024 - FEB 2026 · BOSTON, MA]

 

[ROLE] Teaching Assistant (CS240: Coding in C)
[RESEARCH] Agent-to-Agent (A2A) Governance & DPR
[EDUCATION] BA in Computer Science · GPA: 3.58
  • [01] Led labs and office hours across 5 semesters; mentored 100+ students in low-level C programming, Linux workflows, memory management, and debugging (pointers, segfaults, file I/O).
  • [02] Authored research on Delegation Provenance Records (DPR), designing a zero-trust governance middleware layer for multi-agent systems with cryptographic delegation chains and runtime scope enforcement.
  • [03] Deployed a voice AI research experiment (n=140, Prolific) on conversational formality in older adults at Northeastern University (CHAI Lab); built real-time research dashboard for CHI/CUI.
  • [04] Evaluated LLMs' ability to resolve data races by integrating domain-specific static analysis tools via MCP; presented at the New England Programming Languages and Systems Symposium (NEPLS) at BU.
  • [05] Honors: AstraZeneca Global Health Hackathon Runner-Up (2nd place) · UMass Boston AI Symposium Hackathon Crowd Favorite ($1,000 prize).
C (LOW-LEVEL) MEMORY MANAGEMENT LINUX A2A GOVERNANCE LEAN 4 / COQ MCP ANALYSIS RESEARCH
+ [SECTION // 03 · TECHNICAL ARSENAL]
CANONICAL SPECIFICATION 04 PILLARS
[SPEC // 01 · SECURITY & RUNTIMES] ENTERPRISE CORE

Runtime Authorization & Multi-Agent Governance

Designing deterministic policy enforcement, runtime boundary controls, and cryptographic delegation for autonomous agent workflows.

  • + OPA / Rego Access Control: Runtime authorization engine integrating intent-deviation detection, SOD violations, and behavioral monitoring (Northern Trust POC).
  • + Model Context Protocol (MCP): Protocol design, static analysis tool evaluation (NEPLS at BU), and 'Agentic50' global SE training.
  • + Zuma AI Security Platform: Product and technical leadership through GA, securing ~$15M in enterprise deal value across 22+ RFPs.
  • + Identity & Auth: JWT-signed pre/post-tool webhooks and Okta to Salesforce / Copilot Studio OAuth 2.0 token exchanges.
OPA / REGO MCP PROTOCOL A2A / H2A SOD JWT WEBHOOKS OKTA OAUTH
[SPEC // 02 · REAL-TIME VOICE AI] STARTUP & PRODUCTION

Low-Latency Conversational Audio & Voice Pipelines

Full-duplex acoustic streaming, real-time interruption handling, and dialectic dialogue engines for high-stakes environments.

  • + Voice Infrastructure (VAPI): Sub-second audio streaming, WebSockets, telephony integration, and acoustic turn-taking.
  • + Prompt Builder & Visual Canvas: Visual node-based workflow editor, transcript-to-flow automation, and version diff tracking at Spoqen.
  • + Dialectic Socratic Learning (Meno): Low-latency voice tutoring platform built for verbal comprehension feedback and natural conversational pacing.
  • + Healthcare AI (Calico Care): HIPAA-compliant voice outreach architecture with 12+ scenario automated clinical safety testing suite.
VAPI WEBSOCKETS FULL DUPLEX HIPAA VOICE VISUAL CANVAS
[SPEC // 03 · LOW-LEVEL SYSTEMS] SYSTEMS ARCHITECTURE

Systems Programming, Memory Architecture & Concurrency

Direct memory management, pointer safety, operating systems fundamentals, and performant concurrent native code.

  • + C (Low-Level Systems): 5 semesters as CS240 TA mentoring 100+ engineers in manual heap allocation (malloc/free), stack vs heap mechanics, pointers, segfaults, and POSIX file I/O.
  • + Rust: High-performance zero-cost abstractions, memory-safe concurrency, and strict borrow-checker safety for cryptographic verification.
  • + Linux & POSIX: Shell automation, IPC, process management, memory profiling, and low-level Linux server administration.
  • + Static Analysis: Research on detecting and resolving concurrency data races via automated static toolchains.
C (POSIX) MEMORY ALLOCATION RUST CONCURRENCY LINUX
[SPEC // 04 · RESEARCH & CLOUD] FRONTIER SPEC

Formal Verification & Multi-Cloud Agent Infrastructure

Theorem automation, cryptographic delegation research, and enterprise foundation model gateway orchestration.

  • + Theorem Automation Research: LLM-assisted formal proof generation and proof verification workflow using Lean 4 and Coq interactive theorem provers.
  • + Delegation Provenance Records (DPR): Zero-trust governance middleware research for multi-agent systems with cryptographic delegation chains.
  • + Enterprise Cloud Gateways: Multi-cloud agent discovery and guardrails across AWS Bedrock, GCP Vertex AI, Azure AI Foundry, and Copilot Studio.
  • + Data & Infrastructure: Docker containerization, Redis state caching, Supabase Postgres with Row-Level Security (RLS), and Tableau KPI telemetry.
LEAN 4 / COQ DPR RESEARCH AWS BEDROCK GCP VERTEX AZURE FOUNDRY
+ [SECTION // 04 · AWARDS & RESEARCH]
CANONICAL RECORD: 2024 - PRESENT STACKED TELEMETRY
/ AWARDS & RESEARCH

Peer-reviewed dispatches & competitive AI hackathon victories:

/007
  • S/001

    AstraZeneca Global Health Hackathon

    Early-Onset Alzheimer’s Acoustic Biomarker Agent Team Lead & Voice AI Architect · Global Competition

    Led an engineering team to architect an end-to-end voice AI agent diagnosing subtle linguistic hesitations, conversational pause variance, and acoustic vocal biomarkers associated with early-stage Alzheimer’s in older adult cohorts.

    [LATENCY] Sub-second conversational turn-taking & continuous acoustic streaming
    [CLINICAL] Acoustic vocal biomarker extraction & real-time hesitation metrics
    [RESULT] Global runner-up among international engineering teams
    [HACKATHON WIN]
    [VERIFIED PILLARS]
    ASTRAZENECA VOCAL BIOMARKERS ALZHEIMER'S AI HEALTHCARE PIPELINE
    [JURY OUTCOME] 2ND PLACE GLOBAL HONORS
  • S/002

    UMass Boston AI Symposium Hackathon

    Real-Time Patient Vocal Distress & Respiratory Classifier Team Lead & Lead Engineer · UMass Boston AI Symposium

    Led the development of a proprietary voice analysis classification model engineered to classify physiological distress and acute respiratory strain in chronic disease patients directly from incoming live conversational audio streams.

    [AWARD] Voted Crowd Favorite & awarded First Prize $1,000 purse
    [MODEL] Custom acoustic feature extraction & low-latency inference classifier
    [DOMAIN] Continuous non-invasive chronic disease patient monitoring
    [FIRST PRIZE]
    [VERIFIED PILLARS]
    $1,000 FIRST PRIZE CROWD FAVORITE ACOUSTIC CLASSIFIER INFERENCE PIPELINE
    [JURY OUTCOME] SYMPOSIUM FIRST PLACE
  • S/003

    Delegation Provenance Records (DPR)

    Zero-Trust Governance Middleware for Multi-Agent Systems Co-Author & Security Researcher · UMass Boston (2026 - Present)

    Co-authoring formal research establishing Delegation Provenance Records (DPR), a zero-trust governance middleware protocol enforcing runtime authorization, cryptographic lineage tracking, and execution boundary validation across autonomous Agent-to-Agent (A2A) chains.

    [SECURITY] Tamper-evident delegation tokens preventing confabulated agent authority
    [INTERCEPT] Deterministic runtime hooks preventing out-of-boundary tool invocations
    [PROTOCOL] Multi-agent Segregation of Duties (SOD) & dynamic scope reduction
    [A2A ZERO-TRUST]
    [VERIFIED PILLARS]
    A2A ZERO-TRUST DPR PROTOCOL CRYPTOGRAPHIC LINEAGE RUNTIME SCOPE
    [DISPATCH STATUS] PAPER IN PROGRESS // 2026
  • S/004

    NEPLS Symposium at Boston University

    Concurrency Data Race Resolution via Model Context Protocol (MCP) Research Speaker · Boston University (Fall 2025)

    Delivered a research talk at the New England Programming Languages and Systems (NEPLS) Symposium, evaluating foundation models' ability to accurately detect and patch concurrency data races by orchestrating domain-specific static analysis tools over MCP.

    [EVALUATION] Assessed LLM patch fidelity against ThreadSanitizer & race detector verifiers
    [PROTOCOL] Live architectural demonstration of static analysis tool execution over MCP
    [CONCURRENCY] POSIX multithreading memory hazard mitigation & lock contention analysis
    [SYMPOSIUM TALK]
    [VERIFIED PILLARS]
    NEPLS TALK DATA RACES MCP INTEGRATION STATIC ANALYSIS
    [DISPATCH STATUS] INVITED SPEAKER // BU
  • S/005

    Northeastern University (CHAI Lab)

    Conversational Voice AI Formality & Pacing in Older Adults HCI Researcher · Northeastern CHAI Lab (Oct 2025 - Present)

    Executing an empirical human-computer interaction research trial (n=140 via Prolific) investigating how conversational formality, acoustic turn-taking pacing, and persona design influence trust, comprehension, and comfort in older adult cohorts.

    [SAMPLE SIZE] n=140 participants recruited and deployed via Prolific platform
    [TELEMETRY] Engineered real-time dashboard tracking p-values, formality scores & audio logs
    [TARGET] ACM CHI / CUI conference poster submission
    [HCI RESEARCH]
    [VERIFIED PILLARS]
    PROLIFIC (N=140) CHAI LAB RESEARCH DASHBOARD CHI / CUI '26
    [DISPATCH STATUS] POSTER SUBMISSION ACTIVE
  • S/006

    Theorem Automation (Lean 4 & Coq)

    LLM-Assisted Interactive Formal Proof Generation & Verification Formal Methods Researcher · UMass Boston (2026 - Present)

    Developing an LLM-assisted formal proof generation and verification pipeline leveraging foundation models to synthesize tactical proof steps deterministically validated by Lean 4 and Coq kernel checkers.

    [KERNEL] Zero-hallucination guarantee through proof-checker kernel validation
    [TACTICS] Automated tactical search across inductive propositions & mathematical lemmas
    [DISPATCH] Selected for presentation at UMass Boston Research Symposium 2026
    [FORMAL METHODS]
    [VERIFIED PILLARS]
    LEAN 4 COQ PROVER TACTICAL PROOFS SYMPOSIUM 2026
    [DISPATCH STATUS] SYMPOSIUM 2026 DISPATCH
  • S/007

    JFK Award Nominee & University Citations

    University Leadership Citations & Academic Merit Fellowships BA in Computer Science · GPA 3.58 / 4.00 · CS240 Teaching Assistant
    [JFK NOMINEE]
    John F. Kennedy (JFK) Award Nominee University-wide recognition for civic dedication, intellectual distinction, and leadership
    [DEAN'S LIST]
    Dean's List (5 Consecutive Semesters) Consistent academic distinction in Computer Science coursework (GPA: 3.58)
    [MERIT SCHOLAR]
    Elizabeth Hanopol Merit Scholarship Competitive university merit fellowship based on academic excellence in computing
    [LEADERSHIP]
    Student Senate Merit Scholarship Awarded for instructional dedication and mentorship across 100+ students as CS240 TA
    [HONORS]
    [DEGREE CONFERRED] BA COMPUTER SCIENCE GPA: 3.58 / 4.00
    [HONORS SUMMARY]
    JFK NOMINEE DEAN'S LIST (5X) HANOPOL SCHOLAR SENATE SCHOLAR

  

Let’s build high-integrity systems together.

Whether you're exploring real-time voice AI pipelines, low-level concurrency engines, zero-trust security middleware, or full-stack products, I'm always open to discussing engineering challenges, research collaborations, or technical opportunities.

[01 // DIRECT LINE]
clark.ohlenbusch@gmail.com +1 (781) 424-3027 Boston, Massachusetts · EST (UTC-5)
[02 // NETWORKS & REPOSITORIES]
github.com/ClarkOhlenbusch ↗ linkedin.com/in/clark-ohlenbusch ↗
[03 // ROLES & RESEARCH]
Engineering Roles & Research (2026) Preferred Focus: Real-Time Systems, Voice AI, Security, Rust/C/TypeScript
CLARK OHLENBUSCH © 2026 // ARCHIVE SPECIFICATION V3.0 · BOSTON, MA