Skip to main content

About Dwayne Helena

At a glance​

SignalProof point
Current roleDirector of AI Engineering at V2 AI
Experience25+ years across enterprise engineering, data, platform, DevSecOps, and technology transformation
Core environmentRegulated enterprise delivery, including banking, financial services, and consulting
AI focusGoverned enterprise AI adoption, AI-powered SDLC, evals, controls, and measurable delivery improvement
Builder proofOpenClaw autonomous agent system, Feature Spec Generator sponsorship, and this public knowledge hub
Conversation fitCTO, VP Engineering, AI engineering, DevSecOps transformation, and responsible GenAI adoption

I am Director of AI Engineering at V2 AI. I help enterprises move from AI experiments to governed production systems, with a focus on how teams build, secure, measure, and improve them.

V2 AI announced my appointment on 24 June 2026 as one of three senior leadership appointments. I bring more than 25 years of experience across technology, data, platforms, DevSecOps, and engineering transformation in banking, financial services, and consulting.

I treat AI and engineering performance as operating model problems, not only technology problems. Tools matter. The practices, controls, feedback loops, and learning systems around them determine whether improvements last.

Alongside my client and leadership work, I build and operate OpenClaw, a personal infrastructure project that runs agent workloads against real automation problems.

How I work​

Early in my career, I learned that many difficult technology problems start with the organisation. A sound architecture still degrades when the teams building and operating it lack shared standards for quality, security, and delivery performance.

I start transformations with the operating model: team structure, measures of success, and the feedback loops connecting delivery outcomes to engineering decisions. Technology choices follow. Maturity models, DORA metrics, and capability frameworks make performance visible enough to improve.

Leaders should still build. Sponsoring the Feature Spec Generator, operating OpenClaw, and maintaining this knowledge base keep my decisions grounded in delivery work. This site itself is maintained through Claude Code — the daily Tech Leaders Brief pipeline, the arXiv research sweep, and the site's own content changes all run through it, which is part of why the AI-powered SDLC recommendations here come from direct use, not a vendor comparison exercise.

Selected work​

Enterprise engineering and DevSecOps transformation​

Before joining V2 AI, I held senior engineering leadership roles across highly regulated environments, including enterprise-wide DevSecOps, quality engineering, and platform transformation initiatives at a Tier-1 bank. That work included:

  • Led enterprise-wide DevSecOps and testing transformation across engineering, cyber, risk, and platform teams with different technology stacks and regulatory obligations.
  • Defined DevSecOps maturity metrics for pipeline security, infrastructure as code, automated testing, and change management automation.
  • Worked with business units to resolve data, tooling, and SDLC gaps. Improvement plans accounted for team capacity and technical debt instead of treating a score as the outcome.
  • Set delivery cadence, adoption checkpoints, and executive reporting with the DevSecOps Transformation Office. The reporting tracked whether adoption improved delivery outcomes.
  • Helped embed cyber controls in golden paths and pipelines, reducing the manual compliance work required from engineers.

AI-powered engineering: the Feature Spec Generator​

I sponsored and guided work from the Seattle Tech Hub to build the Feature Spec Generator. It uses language models to turn short requirements into executable BDD specifications. The aim is to help engineers find ambiguity earlier in the software lifecycle, not replace their judgement.

The generator takes a short feature description and produces Gherkin specifications covering the expected path, edge cases, and error handling. Human engineers review and approve the output before it enters the testing pipeline. That control is deliberate: the system helps with requirements discovery and test design, while engineers remain accountable for the specification.

OpenClaw: autonomous agent system​

OpenClaw is an autonomous agent system that I designed, built, and operate on my own infrastructure. It began as an experiment and now handles home and development automation across several models and services:

  • Routes tasks across Claude Opus 4.6, Claude Sonnet 4.6, and local Ollama models using cost-aware rules.
  • Runs 88 scheduled automation tasks for security monitoring, geopolitical intelligence, portfolio management, home energy, sleep scheduling, and infrastructure health.
  • Operates more than 10 intelligence skills that scan feeds, APIs, and news sources and send source-linked briefs.
  • Integrates with Home Assistant for energy scheduling, sleep cooling, camera analysis, and more than 140 device automations.
  • Provides more than 140 installed skills across intelligence, finance, home automation, development, communication, and security.
  • Includes doctor --fix diagnostics for common infrastructure faults on a real-time Linux kernel.

The Living Architecture documents OpenClaw in detail. The AI Research section connects those implementation choices to the research behind them.

Leadership focus​

  • Governed enterprise AI adoption with measurable changes to delivery, assurance, and operations.
  • Resilience, performance engineering, and automated quality gates that make quality visible and systemic.
  • Incremental DevSecOps, secure SDLC, and testing modernisation for regulated environments.
  • Governance frameworks, risk assessments, and adoption playbooks for AI-enabled engineering workflows.
  • Career pathways for engineering, quality, and platform roles.

Topics I work on​

My current areas of interest are:

  • CTO and VP Engineering leadership in regulated organisations.
  • Enterprise AI engineering with governance, human review, evidence, controls, and accountability.
  • DevSecOps operating models, metrics, and engineering culture.
  • Responsible use of generative AI in engineering workflows.
  • Resilient engineering organisations and team design.
  • Agent systems, safety controls, and autonomous infrastructure.

References​

  • Forsgren, N., Humble, J., & Kim, G. (2018). Accelerate: The Science of Lean Software and DevOps. IT Revolution Press.
  • Kim, G., Humble, J., Debois, P., & Willis, J. (2016). The DevOps Handbook. IT Revolution Press.
  • DORA Team. (2023). State of DevOps Report. Google Cloud / DORA. https://dora.dev
  • Humble, J. & Farley, D. (2010). Continuous Delivery: Reliable Software Releases through Build, Test, and Deployment Automation. Addison-Wesley.