Doug Macnally

Profile

AI/ML engineer with hands-on experience shipping production systems across edge computing, computer vision, signal processing, and LLM integration. I build end-to-end — from training classifiers and deploying inference pipelines to machining the hardware enclosures and fabricating carbon fibre airframes that house them.

Currently building SENTINEL, an AI-driven counter-UAS RF detection platform targeting Canadian defence contracts, and operating multiple AI-powered products including an autonomous trading intelligence system and an AI dental visualization service.

Unique combination of software engineering, ML deployment, embedded systems, and advanced manufacturing. 15+ years CNC machining and composites fabrication. Founder of three incorporated companies under Greystone Capital Holdings. Canadian citizen, fully remote.

Technical Skills

Selected Projects & Products

SENTINEL — AI Counter-UAS Detection Platform

• Designed and architected an AI-driven RF detection system for identifying, classifying, and geolocating unmanned aerial systems using software-defined radio and deep learning
• Built signal classification pipeline: GNU Radio SDR front-end, spectrogram extraction, CNN classifier (TensorFlow Lite), Cursor-on-Target output to ATAK tactical displays
• Engineered multi-node sensor fusion architecture: distributed nodes on Raspberry Pi CM5 with LoRa mesh transport (Reticulum), central fusion on NVIDIA Jetson AGX Orin (275 TOPS)
• Integrated detection pipeline with encrypted mesh network — every detection propagates as CoT XML over LXMF to all field terminals simultaneously
• Targeting CFB Suffield C-UAS Sandbox demonstration (2028) and IDEaS programme funding

SOL Intelligence — Autonomous Trading Platform

• Built and deployed a 43-module ML trading intelligence platform processing cryptocurrency market data with automated feature engineering
• Deployed to DigitalOcean with 3 systemd services, Nginx reverse proxy, SSL, and Streamlit dashboard
• Modules include: FOMC Analyzer, Cycle Position tracker, Market Breadth aggregator, LightGBM predictive model, pattern database builder
• Built Rayne — an LLM analyst module pulling live Supabase context to Qwen 2.5 14B via Ollama for natural language market briefings
• Implemented ICT-style structure analysis: FVG zones, order blocks, liquidity grabs, session timing, volume profile, multi-asset correlation

Greystone Private Network — ZK Mesh Communications

• Designed and deployed a zero-knowledge proof mesh networking system on Reticulum for sovereign encrypted communications
• Built three Groth16 circuits (Circom/snarkjs) and deployed verified smart contracts to Ethereum Sepolia testnet — all proofs under 810 bytes
• Established physical mesh: LoRa32 nodes and T-Deck communicating over encrypted LoRa links, infrastructure-independent

SmileForward — AI Dental Visualization

• Conceived, built, and launched an AI-powered dental visualization product in under 24 hours using ComfyUI, Flux, and custom LoRA models
• Engineered face-preserving inpainting pipeline: differential diffusion maintains facial identity while transforming mouth region only — 30-second renders on RTX 5070
• Three-tier pricing model: consumer, subscription, and dental clinic B2B partnership

PCRL Flying Wing UAV Programme

• Designing blended wing body UAV platform (1,800mm span, MH60 airfoil, ~2.2kg AUW) as SENTINEL sensor carrier
• Built suite of interactive manufacturing tools: Planform Layout, Airfoil Plotter, Rib Template Generator — all producing 1:1 scale SVG/DXF for CNC cutting
• Full carbon fibre composite construction using in-house fabrication capability

AI Automation & Web Development

• Built Make.com automation pipelines: RSS to Claude API content analysis to Notion database
• Deployed production web applications on Netlify with Supabase backends and custom React dashboards
• Created multi-model AI workflows integrating Claude API, Ollama local inference, and ComfyUI image generation

Professional Experience

CNC Operator / Composites Manufacturer

• 15+ years operating multi-axis CNC equipment in precision manufacturing environments
• Self-taught composites fabrication: carbon fibre layup, vacuum bagging, resin infusion, mould design
• Designed and prototyped Pin Gage Depth Verification Jig solving workplace QC problem — priced for commercialization
• CAD design in Onshape: mechanical assemblies, jigs, fixtures, production tooling
• Developing ML-augmented CNC lathe conversion: Jetson Nano computer vision for tool wear detection, vibration monitoring

Companies

Domain Expertise

Education

Self-directed — extensive applied learning in AI/ML, embedded systems, aerospace engineering, blockchain, and defence technology. Portfolio of shipped products demonstrates practical competency across all listed domains.