Paul Richardson
Teacher | AI Engineer | Cloud Architect
Maine, USA
https://www.oddlytrue.ai
paul@oddlytrue.ai
Professional Summary
Experienced technologist, educator, and builder with deep expertise in AI systems, cloud architecture, and translating complex ideas into practical, real-world solutions.
Specializes in designing and deploying end-to-end systems — from AI-powered homes and IoT networks to production-ready cloud architectures. Passionate about teaching and enabling others to build confidently with modern technology.
Core Skills
- Cloud Platforms: AWS (Lambda, S3, DynamoDB, API Gateway, CloudFront, VPC)
- AI / ML: LLM integration, Amazon Bedrock, local AI models (LLaMA), RAG systems
- Programming: JavaScript (Node.js), Python, Java, .NET
- Frontend & UX: React, Vite, modern web development
- Backend & DevOps: Flask, Node APIs, serverless, Docker, CI/CD, Cloudflare, Vercel
- IoT & Hardware: Raspberry Pi, XBee networks, sensor integration
- Teaching & Education: Curriculum design, hands-on instruction, technical mentoring
Teaching Experience
Southern Maine Community College (SMCC)
Instructor – Architectural and Engineering Design
- Deliver courses in Architectural and Engineering Design with strong integration of modern technology
- Designed and implemented hands-on curriculum combining traditional design principles with cloud and AI tools
- Integrate AWS architecture, AI fundamentals, and real-world engineering practices into coursework
- Guide students through practical, project-based learning that bridges design, engineering, and emerging technologies
Additional Instruction
- AWS Developer & Cloud Architecture Courses
- Hands-on training in secure AWS design, serverless systems, and deployment pipelines
- AI Foundations Course
- Designed for beginner to intermediate learners
- Covers AI fundamentals, prompt engineering, local models, and edge computing
- Includes real-world projects using Raspberry Pi and integrated hardware systems
Featured Projects
Halo Lumin
Frictionless file-to-link sharing platform
- Built a platform that transforms text and files into clean, shareable web pages
- Designed for zero-friction sharing: Create → Link → Share
- Enables users to replace traditional file attachments with simple, persistent links
- Stack: Next.js, Supabase, Vercel
AI-Powered Passive Solar Home
Intelligent, locally-run smart home system
- Designed and built a passive solar home with real-time AI monitoring and control
- Tracks temperature, humidity, energy usage, occupancy, and security signals
- Integrated XBee sensor network, Raspberry Pi data layer, and high-performance local AI node
- Focused on optimizing energy efficiency and comfort through continuous learning
Tokenizer 5000 (RatStack)
Adaptive, local-first AI learning system
- Developed agentic workflows using retrieval-augmented generation
- Designed for continuous improvement using real-world data streams
- Emphasizes efficient, local model execution and adaptive system behavior
FELIX – Veterinary AI Assistant
AI-powered diagnostic support system
- Built to analyze symptoms and assist veterinarians with diagnostic insights
- Designed with a secure, federated data-sharing model for clinical collaboration
- Focused on improving decision-making for less experienced practitioners
Approach & Philosophy
- Build first, refine second
- Keep systems simple, secure, and explainable
- Prefer local-first AI whenever practical
- Design for real users, not just other engineers
- Focus on practical outcomes and clear communication
Select Strengths
- Explains complex technical concepts clearly (teacher-first mindset)
- Bridges hardware, software, cloud, and AI seamlessly
- Rapid prototyping and iterative development
- Practical, results-oriented problem solving
Interests
- Agent-based AI systems and architectures
- IoT and intelligent environments
- Energy-efficient and passive solar design
- Teaching and mentoring future technologists
- Building tools that people genuinely enjoy using
Let's Connect
Interested in collaborating, teaching opportunities, or building something meaningful together?
Send me a Halo at https://www.halolumin.io
Or reach out directly: paul@oddlytrue.ai