About
Hi! My name is Nguuma, and I'm a Senior AI Engineer, Application Security Specialist, and open-source researcher. My expertise sits at the intersection of software, LLM engineering, and cybersecurity designing, building, and hardening secure systems across FinTech, healthcare, public sector, broadcasting, and security research. Over the past decade I have worked with organisations that cannot afford to get AI wrong regulated industries, vulnerable populations, and critical infrastructure and I have built systems they trust.
I hold an MSc in Computer Networks and Security from the University of Essex, and I am a CISSP-certified security professional one of the most respected credentials in the field. I have worked across some of the most impactful environments in AI and security. In FinTech, I have built an AI-powered analysis and fraud detection system, and designed and shipped a FinTech LLM application end-to-end that cuts broadcast episode production from 5–8 hours down to 3–4 minutes, with a full SAST audit returning zero high-severity findings across 2,777 lines of code. Before that, I was Technical Lead for PadiChat at RNW Media a private AI counselling platform I built from the first commit to production for young Nigerians seeking sexual and reproductive health support, holding 99.9% uptime and automating over 1,000+ routine counselling inquiries.
I have published three open-source fine-tuned small language models on Hugging Face, each built to run inside an organisation's own infrastructure with no sensitive data leaving the environment. My Security-SLM models covering the full AI attack surface from SOC triage to agentic lateral movement and RAG vector-store poisoning have surpassed 3,000+ combined downloads. I also publish Med-Lab-finetuned, applying the same methodology to clinical laboratory diagnostic reasoning for resource-constrained environments.
Aside from building, I am a sought-after voice on AI security and sovereign AI adoption. I have been cited as an expert in The Guardian Nigeria, Daily Trust, Radio Nigeria, Leadership Nigeria, and AllAfrica. I have spoken at GDG Build with AI 2026 and World Creativity and Innovation Day 2026, and I serve as an Independent Research Affiliate (AI/LLM Application Security R&D) at Entrick Information Systems conducting applied research into emerging LLM attack surface and secure AI adoption for regulated environments.
My career ambition is to be the engineer and researcher who proves that security and AI capability are not in tension that the most powerful systems are also the safest ones. I want to shape the standards by which the world's most sensitive AI deployments are built, and leave behind models, tools, and frameworks that make secure AI accessible to the organisations that need it most.
How I got here
Built the Financial AI Podcast Generator end-to-end from data ingestion to audio output cutting eToro episode production from 5–8 hours to 3–4 minutes. Led Publish24, a multi-agent content platform built with FastAPI, Django REST, Next.js and LangGraph, reducing content generation time by 70%.
Set the technical direction and built PadiChat from the first codebase to production a private AI counselling platform offering RAG-based answers, mood tracking, appointments and referrals for young Nigerians seeking SRHR support. Deployed with OWASP-aligned guardrails, prompt-injection controls, and GDPR/HIPAA-equivalent data protection across some of the most fragile information environments in the world.
Translated NCSC Cyber Essentials into a plain-language compliance assistant for SME owners, built on a domain-specific LLM fine-tuned with LoRA and Unsloth. Also designed BizCyberQuest an interactive security-simulation game aligned with the NCSC Cyber Assessment Framework, grounded in research covering 1,400+ UK businesses.
Led architecture and security for the Federal High Court electronic affidavit platform across 7 states replacing paper-based workflows with online registration, identity checks, and certified document delivery. Dashboards recorded 79,049 submitted applications, 72,771 approvals, and over £217,327 first-year revenue.
Designed and secured FRCN's online news portal and virtual radio infrastructure including core servers, Juniper firewalls, and Azure environments helping expand digital broadcasting across 7 state capitals for Africa's largest public radio network.
Credentials & affiliations
Media & speaking
Quoted on Nigeria's sovereign AI strategy, low-infrastructure AI systems for critical sectors, and local model deployment in healthcare, finance, and government.
Delivered a technical session on building secure AI agents in Golang. The session was selected and later published on YouTube.
Invited to speak to founders, SME operators, and public leaders on practical AI use in business workflows.
Presented PadiChat at the Digital Media Impact Summit 2024 winner announcement after the platform won the RNW Media Innovation Fund, and gave subsequent stakeholder updates.
Invited by the national public broadcaster to discuss how locally adapted AI surveillance could support work against insecurity in Nigeria.
Published comments on AI surveillance, predictive analytics, and national security policy in Nigeria.