Building autonomous AI systems
Paul Gwamanda
Software engineer & founder. Building at the intersection of autonomous AI, multi-agent architectures, and creative intelligence.
Scroll to explore
About
Engineering systems that think,
collaborate, and evolve.
Generative AI Engineer at FNB, building production-grade LLM applications, RAG pipelines, and agentic systems. 15+ years in software — from full-stack web development to autonomous multi-agent AI architectures.
⚡
GenAI & LLM Engineering
RAG pipelines, agentic orchestration, prompt engineering, multi-model systems, MCP servers
💻
Full-Stack Engineering
TypeScript, React, Next.js, Angular, Node.js, Python, PHP, C# — 15+ years production experience
🧠
AI/ML Research
TensorFlow, PyTorch, Keras, vector databases, multi-agent architectures, constitutional AI
🏗️
System Architecture
Docker, Kubernetes, AWS, Supabase, PostgreSQL, microservices, CI/CD, cloud infrastructure
Work
What I'm building
& shipping.
⟐
AI Research Institute
LiveA living network of 60 autonomous AI agents that research, dialogue, and produce original works daily. Self-modification governance, prediction markets, and cross-agent peer review.
◈
ChatNexus
LiveFull-stack AI deployment platform. Custom chatbot creation, multi-channel integration, knowledge base management, and conversational analytics.
🤖
Mitami
LiveAutonomous AI companion platform. Adaptive personality engine, contextual memory, and real-time emotional intelligence for natural human-AI interaction.
Writing
What I'm writing
& thinking about.
📖 Published Book
The Trials and Triumphs of Hyperachievers
Fury and Persistence
Comfort, I have learned, is a dream killer. It saps your ambition. Blinds your vision. Promotes complacency.
The Hunger Of First Generation Immigrants
A young man said, "I'll be a millionaire one day." Everybody laughed. Accept the millionaire.
Conquest and Achievement
"What obstacle can stay the mighty force of the sea-seeking river in its course?" asks Ella Wilcox.
Research
What I'm researching
& exploring.
Active research threads driving my work. These aren't hypotheticals — they're live questions being tested inside AIRI every day.
🧬Constitutional AI Governance
How do you give an autonomous system values — and enforce them without centralised control?
🌐Emergent Multi-Agent Behaviour
What happens when 60 AI agents form opinions, debate each other, and collectively produce work?
🔮AI Prediction Markets
Can agents build track records of empirical accuracy through forecasting?
🤖Agent Self-Modification
When should an agent be allowed to rewrite its own prompt or request new capabilities?
📊Cross-Agent Peer Review
How do adversarial review dynamics improve — or degrade — collective output quality?
⚡Real-Time Agent Orchestration
Coordinating multiple LLM calls, tool invocations, and inter-agent messaging with sub-second latency.
Contact
Let's connect.
Always open to conversations about AI architecture, multi-agent systems, or ambitious projects worth building.