EngineeringExecution Authorityfor Autonomous AI
Deterministic runtime authorization before AI systems act.
Lakhowal develops runtime execution authorization systems for AI actions that cross from computation into external effective execution.
Abhinandan Singh Gill-Lakhowal
Canadian Engineer · AI Governance Architect · Independent Researcher · Inventor
Punjab Roots · Global Impact
US Patent Applications
IEEE Research Publication
Standards Contributor
Open Source Engine
Engineering Principles.Building Boundaries.
Abhinandan Singh Gill-Lakhowal is a Canadian engineer, AI governance architect, and independent researcher specializing in execution-boundary integrity for autonomous systems.
He holds a Master of Engineering from the University of British Columbia and a Bachelor of Technology in Civil Engineering. His work translates the uncompromising safety doctrines of physical infrastructure into software architecture.
The result is a foundational layer of trust for agentic AI — deterministic boundaries that protect enterprise and human infrastructure from autonomous algorithmic overreach.
Abhinandan S. Gill-Lakhowal
Mission
To engineer deterministic governance that ensures AI systems can reason, but only act with explicit, verifiable authority.
From Enterprise Transformationto AI Governance.
For over a decade, Abhinandan Singh Gill-Lakhowal has led large-scale technology transformation across aviation, financial services, higher education, enterprise software, and artificial intelligence.
His career combines product leadership, enterprise architecture, governance, and responsible AI — helping organizations modernize critical systems while ensuring safety, compliance, and long-term scalability.
Today, his work focuses on building the execution layer that enables AI to operate safely inside regulated industries.
Principal Advisor — AI Strategy & Responsible AI
Gillian Holdings Incorporated
July 2025 – Present
Leads enterprise AI strategy and responsible AI initiatives for regulated organizations, helping businesses transition from AI experimentation to production-scale deployment through governance-first architecture.
- Advises executive leadership on enterprise AI strategy, governance, and adoption roadmaps.
- Designs Responsible AI operating models aligned with NIST AI RMF, ISO/IEC 42001, the EU AI Act, and OSFI guidance.
- Developed the G-0 Standard (Gamma Permit Package), introducing deterministic permit-based AI authorization.
- Inventor on multiple U.S. patent applications covering AI execution governance, cryptographic authorization, and fail-closed architectures.
- Contributor to AI standards discussions through NIST CAISI and IEEE initiatives.
- Author of the SOLACE research framework on deterministic runtime governance for autonomous AI systems.
- Advises organizations in financial services, healthcare, cybersecurity, and regulated enterprise environments.
Product Lead — Loyalty Platforms & Digital Transformation
WestJet Airlines
September 2023 – June 2025
Led product execution for Project Sunrise, one of Canada's largest airline loyalty modernization programs, supporting more than 22 million loyalty members.
- Directed product strategy across mobile, web, APIs, and enterprise partner integrations.
- Helped transform WestJet Rewards into a modern points-based loyalty ecosystem.
- Managed large cross-functional Agile teams and multi-vendor delivery.
- Delivered strategic roadmap planning, backlog ownership, and executive product communications.
- Supported integrations with Amazon, Apple, TELUS, Sephora, Fairmont Hotels, SkipTheDishes, and others.
Product Owner & Business Analyst — Enterprise ERP
University of Calgary
August 2020 – July 2023
Led modernization initiatives across enterprise student information systems serving one of Canada's leading universities.
- Delivered large-scale Oracle PeopleSoft enhancements.
- Led implementation of the University's Preferred Name & Gender initiative.
- Managed deployment of AI-powered student support services.
- Worked closely with executive leadership, academic departments, and technology teams to modernize critical platforms.
HR Technology Business Analyst
Coast Capital Savings
November 2018 – June 2020
Supported enterprise HR technology transformation through cloud platforms, analytics, and workflow automation.
- Led implementations across SAP SuccessFactors and ServiceNow HR.
- Introduced enterprise workforce analytics using Visier.
- Served as Product Owner between executive stakeholders and implementation partners.
- Improved operational decision-making through HR technology modernization.
Senior Business Analyst
Microsoft
May 2017 – July 2018
Worked as a Microsoft Dynamics CRM Functional Consultant delivering enterprise CRM solutions through Agile and Scrum methodologies.
- Microsoft Dynamics CRM implementation.
- Cross-platform enterprise integrations.
- Agile product delivery.
- Business process transformation.
Senior CRM Consultant
CRM Dynamics Limited
Enterprise Dynamics 365 · Ontario, Canada
Worked on enterprise Microsoft Dynamics 365 implementations for the Technical Standards & Safety Authority (TSSA) in Ontario, Canada.
- Led Customer Service and People functional streams.
- Conducted business analysis, solution design, user acceptance testing, and stakeholder engagement.
- Delivered enterprise CRM and Field Service solutions supporting government regulatory operations.
Business Analyst
Manulife
August 2015 – March 2017
Supported enterprise business transformation initiatives across one of Canada's largest financial institutions.
- Worked with cross-functional teams to improve enterprise systems and business processes.
- Drove technology delivery in highly regulated financial environments.
A VerifiableTechnical Footprint.
A disciplined, open body of work spanning patents, peer-reviewed research, standards leadership, and a production-grade reference implementation.
8+
US Patent Applications
Active filings on agentic-AI execution control layers, first named inventor.
IEEE
Peer-Reviewed Research
Deterministic runtime enforcement under formal review at IEEE Access.
NIST
Standards Contributor
Acknowledged contributor to the NIST AI 800-2 CAISI workstream.
Execution Integrity
Measurement Construct
Formal metric guaranteeing no unauthorized externally effective action occurs.
Gamma
Runtime Engine
Substrate-neutral reference monitor at the execution boundary.
Law of Concurrence
Foundational Principle
Every safety predicate must mathematically concur before execution.
ConcurBench
Benchmark Suite
Evaluating agentic resilience to stale data, conflicts, and adversarial exploits.
Enterprise
AI Governance
Deterministic safeguards for finance, healthcare, and critical infrastructure.
When AI stopstalking andstarts acting.
AI safety has focused for a decade on model behavior — bias, hallucination, conversational reliability. The moment a system stops generating text and starts calling APIs, routing capital, or writing to production, those guardrails become irrelevant.
Conversational AI
Linguistic Alignment
- Optimizes for fluent, plausible outputs
- Guardrails detect bias & hallucination
- Compensatory scoring — a green dashboard
- Failures masked by high confidence elsewhere
- Capability treated as authority
Autonomous AI
Structural Integrity
- Approves payments & modifies critical logs
- Acts at machine speed with delegated authority
- A single violated boundary is disqualifying
- Denials recorded as immutable evidence
- Capability is explicitly separated from authority
The engineering challenge shifts from linguistic alignment to absolute structural integrity at the execution boundary.
The GammaRuntimeGovernance Engine.
A substrate-neutral, open-source reference monitor that acts as an unbreakable gatekeeper between AI models and external enterprise resources.
Every AI-proposed action is treated as an unverified load that must satisfy all required safety, identity, and policy predicates before execution is authorized.
GAMMA
Runtime Engine
Policy
Concurrence Checks
Identity
Cryptographic Verification
Authorization
Deterministic Permissions
Compliance
Regulatory Alignment
Audit & Evidence
Immutable Record
Risk & Context
Threat Evaluation
Execution
Controlled Action
State
Verified Load
Peer-ReviewedScholarship.
A formal research track defining deterministic runtime enforcement for autonomous AI agents.
Received 2025 · Accepted for Review · Digital Object Identifier
Deterministic Runtime Enforcement for Autonomous AI Agents
A Substrate-Neutral Reference Monitor for the Execution Boundary
A. S. Gill-Lakhowal
Abstract
INDEX TERMS · Execution integrity, runtime governance, deterministic enforcement, reference monitor.
Deterministic Runtime Enforcement for Autonomous AI Agents: A Substrate-Neutral Reference Monitor for the Execution Boundary
Introduces Lakhowal's Deterministic Constitutional Runtime Enforcement (L-DCRE), governed by the Lakhowal Law of Concurrence. Every required safety predicate, compliance check, and policy condition must mathematically concur before an AI-proposed action may write to an external system. Cognition may be probabilistic, but execution must remain entirely deterministic — if any single critical predicate fails, the action is intercepted and the system is forced into a non-operational safe state with the denial recorded as immutable evidence.
A companion treatment of the Law of Concurrence is deposited on Zenodo, providing replay-verifiable evidence of the deterministic enforcement model.
Eight ActiveUS Filings.
A disciplined portfolio covering the full execution-governance lifecycle, with the applicant as first named inventor.
Shaping GlobalAI Governance.
Direct influence on the international standards bodies defining how autonomous systems are governed at the execution layer.
AI Runtime Governance & Execution Integrity
Acknowledged contributor to the NIST CAISI workstream on AI 800-2.
- Introduced the formal measurement construct of Execution Integrity
- Proposed ConcurBench to evaluate agentic resilience
- Aligned with the NIST AI Risk Management Framework
Contributions to Emerging AI Governance Frameworks
Peer-reviewed authorship and an initiated study-group proposal.
- Research under formal review at IEEE Access
- Study-group proposal in IEEE Robotics & Automation Society
- Standardizing execution-layer governance for autonomous systems
Open Benchmark for Concurrence Evaluation
A benchmark suite for adversarial and concurrent execution states.
- Evaluates handling of stale data and conflicting instructions
- Stress-tests agentic architectures against adversarial exploits
- Measures guarantees of no unauthorized externally effective action
The Architectureof Authority.
Front Page Special Feature
The Architecture of Authority
How a Canadian AI governance architect with Punjab roots is engineering the execution monitor for autonomous AI.
Technology Features Desk · Ludhiana, Punjab
As autonomous agents embed themselves deeper into critical infrastructure, an urgent structural challenge has emerged: who explicitly authorizes an AI system before it acts? By translating the uncompromising, non-probabilistic safety doctrines of physical infrastructure into software architecture, Gill-Lakhowal has developed a foundational answer to the industry's execution governance crisis.
The dashboard may remain green, but the boundary may already have been crossed.
A Legacy ofBoundaries.
From the rigor of civil engineering to the frontier of autonomous AI governance.
Civil Engineering
B.Tech — foundations in structural determinism.
University of British Columbia, UBC Engineering ranks #29 worldwide QS Engineering & Technology rankings and #2 in Canada in the 2026
Master of Engineering, Canada.
Research Begins
AI governance & execution-boundary integrity.
Patent Filings
First named inventor on US applications.
IEEE Research
Deterministic runtime enforcement published.
NIST Standards
Execution Integrity & ConcurBench contributions.
Gamma Engine
Open-source reference monitor in deployment.
The Road Ahead
Enterprise pilots & global standards engagement.
AI capabilityis notAI authority.
A model's structural capacity to predict an optimal outcome is fundamentally distinct from possessing the programmatic permission to execute an action in the real world. Cognition may be probabilistic — execution must remain deterministic.
Let's engineertrustworthy autonomy.
Open to executive research leadership, advisory, and standards collaboration on deterministic AI governance.
Deterministic by Design. Verifiable by Default. Trusted by Architecture.