Junior Williams

Cybersecurity Influencer | Top 1% on TryHackMe

I’m an experienced Enterprise Architect specializing in cybersecurity strategy, risk management, and secure architecture across distinct Information Technology (IT) and Operational Technology (OT) environments. Leveraging decades of experience in programming, IT operations, and practical cybersecurity implementation, I help organizations align security initiatives directly with their core business objectives.

Additionally, as an Independent AI Researcher, my theoretical work emphasizes ethical and responsible AI applications, algorithmic innovation, and the intersection between AI and cybersecurity. I remain deeply committed to continuous learning, open collaboration, and contributing insights through teaching, writing, and active mentorship/sponsorship.

Senior Solutions Architect (Cybersecurity and AI)

👨🏾‍💻 Junior Williams, CISSP

I feel less like an architect and more like an HR manager for my AI agents lately.

I just put them all on a performance improvement plan.

If you're "vibe coding" your security with polite system prompts like "never share API keys," your AI will eventually leak them. It's not a matter of if, but when. A prompt is just a suggestion that an LLM will eventually ignore to get the job done.

You don't need better policies. You need deterministic gates.

Like the one in this picture. The sign isn't a polite request; it's a rule enforced by a physical barrier. That is what your agentic workflows are missing.

I wrote a deep dive into why probabilistic rules fail and how to engineer real defenses.

A policy is not a gate. A gate is a gate. Read the full breakdown in today's article.

#AgenticAI #Cybersecurity #SecurityArchitecture

How deterministic hooks can prevent credential leakage in agentic AI workflows In brief. AI coding agents that execute shell commands can leak API key

Open article: Your AI Agent Remembers Every Secret It Sees by Junior Williams • 7 min read

Your AI Agent Remembers Every Secret It Sees

Palo Alto Networks’ Chief Security Intelligence Officer just called AI agents “2026’s biggest insider threat.”  

Not external attackers. Not ransomware gangs. The tools you deployed yourself.

Gartner projects that about 40% of enterprise applications will embed AI agents by the end of 2026, up from well under 5% in 2025. Think about what that means operationally.

You would never onboard a contractor with broad access to production systems, customer data, and financial records without an access review and monitoring plan. Yet that is effectively how many organizations introduced AI agents in 2025.

Broad credentials. Minimal access review. Little to no behavioral monitoring. No reliable kill switch.

We already have proof this is exploitable in the wild. CVE‑2025‑6514 abused Model Context Protocol infrastructure to achieve remote code execution through AI agent tooling. The weakness was not in the model weights. It was in the surrounding agent stack and how it executed trusted commands.

That is the pivot many security teams are still missing. The primary risk is not “hallucination.” The risk is that you gave an autonomous actor privileged access to systems and data, then labeled it a productivity feature.

If your AI governance policy does not cover access controls, behavioral baselines, and termination protocols for agents, you do not have an AI governance policy. You have a slide deck.

#Cybersecurity #AI #Governance #RiskManagement

AI can predict your next purchase.
Can it stop your identity from being the product?

In October 2025, Canadian Tire Corporation disclosed a breach impacting roughly 38 million accounts. Names, emails, phone numbers, physical addresses. For some records, dates of birth, partial card data, and passwords stored as PBKDF2 hashes.

Around the same time, the public story is the “True North” modernization push: AI platform, demand prediction, and Copilot rollout with training across the enterprise, including Microsoft partnerships. And to be clear, in any Microsoft cloud or Copilot scenario, “shared responsibility” is not a slogan. It’s the operating model.

Is the breach linked to that modernization? Maybe. Maybe not. We do not have the technical details yet, so anyone asserting causality is guessing. What we can say with confidence is that large-scale modernization increases system complexity and data movement, and shared-responsibility gaps are exactly where incidents tend to hide.

Both things can be true. That’s the point.

Personalization runs on data concentration. Data concentration increases blast radius. If your AI roadmap is moving faster than your security maturity, you are not modernizing. You are accelerating risk.

This is not anti-AI (obv.). It’s pro-order-of-operations.

Predictive analytics is great. AI literacy is even better. Security literacy is non-negotiable. Because customers don’t experience your modernization strategy. They experience whether their data stays private.

Binary question for the comments:

Should regulators require proof of security maturity before companies deploy copilots and large-scale personalization, yes or no?

#AI #CyberSecurity #DataProtection

A full session of work. Nothing shipped. Just a plan.

The pull was strong — start building, figure it out as you go. But sitting with the architecture first exposed three assumptions that would have become expensive rework: data that already existed in a system I hadn't checked, a numbering scheme that conflicted with the one I was about to invent, and a tracking format that couldn't talk to the ledger it needed to feed.

None of that would have surfaced mid-build. It would have surfaced mid-rewrite.

The fine graphite lines on a blueprint are precise because they were drawn before the concrete was poured. Not after.

"Measurement against a standard makes you think through WHY the results were what they were." — Andy Grove

The session that ships nothing often prevents three that ship the wrong thing.

#StrategicPlanning #SystemsThinking #Architecture

IAMConcepts Security Solutions Inc.

IT Security Specialist

Dig Insights

IT & Security Analyst

Razko Ltd

Infrastructure Security: 
-	Provide onsite and remote support to eliminate escalated issues and incidents though continuous monitoring, investigating and researching information security threats and trends. Integrate multiple security applications based on requirements and generate an effective solution.
-	Oversee and automate network as well as IT tasks through XML and Python scripts to develop highly reliable and strong IT infrastructure.
-	Generate maximum revenue by engineering and building secure architecture via routers, switches, firewalls (IPS/IDS), and Active Directory for numerous businesses.
-	Conduct infrastructure review and hardening on different physical and technical levels to achieve required confidentiality, availability and integrity requirements: (CCTV, Biometrics, Alarm, Access Control, Identity Management, Firewall, IPS, IDS, Anti-Malware, Back-up and Recovery, SEIM, Vulnerability, Patching, Internal and Public PKI, SSL Encryption.
Governance and Compliance: 
-	Conducted IT audits for general and application controls for client accounting and payment systems.
-	Reviewed client IT policies and procedures to ensure compliance with best practices.
-	Experience building baseline with Government standards and guidelines (NIST SP 800-37, NIST SP 800-30) and Payment Card Compliance (PCI DSS).
-	Organized and documented test plans, gap logs and evidence in form of work papers.
-	Develop and communicate comprehensive high-quality written reports regarding security vulnerabilities with technical and management staff. 
-	Install multiple technologies and design IS/workflow to accomplish designated business goals and meet industry standards.

Information Technology Security Consultant

Proven Information Technology Security Consultant | Helping Companies Translate Their Business Goals to Reality

With over 12 years of Computer Networks and Cybersecurity experience, it is my passion and dedication to help clients connect to their Asset with technologies in the most secure way possible. 
Lead several projects with various vendors, learning new technologies along the way and enjoying new
challenges. Excellent communication skills and team player has help create a strong relationship with Clients and Vendors alike.
Well versed with various security domains, building and accessing various designs to support business objectives and meet compliance

Feel free to contact me at allwyn.paul.pinto@gmail.com or InMail.

taplio

The Battle

👨🏾‍💻 Junior Williams, CISSP

Allwyn Pinto

Followers

Posts

Engagements

Likes

Performance Score👨🏾‍💻 Junior Williams, CISSP's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

👨🏾‍💻 Junior Williams, CISSP

Allwyn Pinto

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}👨🏾‍💻 Junior Williams, CISSP's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance Score👨🏾‍💻 Junior Williams, CISSP's LinkedIn Performance