AI infrastructure that ships.
Built to last, handed off ready to run.
I come into funded companies that need senior technical leadership without a permanent hire, build or fix the AI infrastructure and platform layer, and hand over a system that runs. Available for engagements of weeks to months, contracted through Wisdom Engineers. At EVERYANGLE, I took a computer vision SaaS from 0 to €1M+ ARR, raised €3M+ in seed funding, cut infrastructure costs by 80-90%, and maintained 99.99% uptime over five years. At Tendr, I shipped an AI inference engine, a multi-tenant B2B SaaS platform, and a data acquisition pipeline in six months from a non-functional proof of concept. Recognised in the CTO Craft 100.
My Technical Expertise
Hands-on delivery across the AI infrastructure and platform stack.
AI Infrastructure
- Ollama
- llama.cpp
- vLLM
- Unsloth
- Triton Inference Server
- ONNX
- TensorRT-LLM
- SGLang
AI Orchestration
- MCP
- Haystack
- Tool Use
- Function Calling
- Multi-Agent Systems
- Prompt Engineering
- Agentic Patterns
- Context Management
AI Data & Retrieval
- pgvector
- Elasticsearch
- Tesseract
- Vector Embeddings
- Semantic Search
- Knowledge Graphs
- RAG Pipelines
- Document Parsing
AI/ML Integration
- OpenCV
- PyTorch
- TensorFlow
- ONNX
- Triton Inference Server
- NumPy
- Pandas
- SciPy
Cloud and Infrastructure
- AWS
- GCP
- Docker
- Terraform
- Kubernetes
- Helm
- Linux
- Podman
Distributed Systems
- Microservices
- Event-Driven Architecture
- REST APIs
- GraphQL
- gRPC
- Message Queues
- Pub/Sub Patterns
- Circuit Breakers
Data Processing
- ETL/ELT
- Batch Processing
- Stream Processing
- Data Warehousing
- Data Lakes
- Data Modeling
- Processing Patterns
- Data Quality
Back-End Engineering
- Python
- FastAPI
- Flask
- Django
- Bottle
- Celery
- sched
- asyncio
DevOps
- Git
- Bash
- Prometheus
- Elasticsearch
- CI/CD
- Filebeat
- Kibana
- Grafana
Database and Caching
- MySQL
- MariaDB
- PostgreSQL
- SQLite
- Redis
- Valkey
- Schema Design
- Query Optimisation
Code Quality
- pytest
- mypy
- black
- ruff
- isort
- pre-commit
- mkdocs
- Hypothesis
Why Work With Me
What funded companies get when they bring in fractional leadership instead of a permanent hire.
Measurable Business Impact
Proven business track record raising €3M+ in venture capital, building €1M+ ARR software as a service products with 60%+ margins, and leading companies from concept to enterprise scale with intentional and aggressive cost discipline.
Intentional Leadership & Culture
Built and scaled high-performing, multinational, fully remote teams with a 90%+ retention rate. I come in, raise the engineering bar, establish clear standards and practices, and hand over a team that runs without me.
Proven Technical Competency
Over 11 years of hands-on experience building enterprise-scale systems with 99.9% SLA. I'm an expert in cloud-native architecture, distributed systems, AI/ML integrations, and DevSecOps practices that solve real problems.
Selected Engagements
Production systems delivered for clients. Built at scale, handed over running.
AI Procurement Intelligence Engine
Local inference AI platform automating Irish public tender discovery for Irish SMEs across a €2 trillion market. Built on Haystack 2 with locally hosted LLMs, with a 16-stage document ingestion pipeline (PDFs, scanned archives, DOCX) feeds a four-pass hierarchical summarisation flow producing machine-comparable capability profiles for semantic matching. No sensitive documents ever leave the client's infrastructure.
Key Impact & Achievements
- Hierarchical summarisation with deterministic context bounding
- Bidirectional matching-ready LLM output (tender + org composites)
- Cosine similarity keypoint deduplication (BGE-M3, threshold 0.93)
- 16-stage document ingestion pipeline with multi-format OCR support
- Local LLM inference with zero data egress to third-party AI providers
- Fingerprint-based composite staleness detection and auto-regeneration
- 13-worker Celery architecture with PostgreSQL advisory lock coordination
- Dead-letter fault tolerance, culprit documents isolated without batch failure
- Python
- Haystack
- Ollama
- Mistral
- BGE-M3
- PostgreSQL
- pgvector
- Celery
- Valkey
- AWS
- Alembic
- Tesseract OCR
- RAG
- Vector Search
Public Procurement B2B SaaS Platform
Multi-tenant B2B SaaS aggregating Irish government procurement data from etenders.gov.ie and enriching it with an AI vector layer for semantic tender discovery. PostgreSQL Row-Level Security enforces tenant isolation at the database level. Cross-tenant data leaks are structurally impossible. Subscription-gated platform delivers filtered matches to organisations with team management, knowledge base, and notification tooling.
Key Impact & Achievements
- Dual-layer multi-tenancy: RLS at DB + application level
- AI-powered semantic tender discovery via vector brain service
- Automated government portal harvesting pipeline
- Async document processing with PDF conversion and S3 storage
- Stripe-integrated subscription billing with dunning and grace periods
- Fine-grained RBAC with 30+ permission strings
- Fail-fast architecture: misconfiguration is a startup error, not a runtime bug
- Full audit trail via soft deletes and structured request-ID logging
- Python
- Flask
- PostgreSQL
- SQLAlchemy
- Celery
- Valkey
- AWS
- Stripe
- Docker
- Vite
- Row-Level Security
- RBAC
- 12-Factor App
- Multi-tenancy
Public Procurement Data Pipeline
Resilient data acquisition pipeline scraping Ireland's national public procurement portal and structuring tender data for downstream analytics. Built in Python with Playwright browser automation, the system ingests three tender subtypes (CFT, DPS, QS) across a full organisational graph, archives versioned documents to S3, and tracks field-level changes with a JSONB audit trail. Thread-safe, non-abusive, rate-limited design.
Key Impact & Achievements
- Field-level change monitoring with full audit trail
- MD5-deduplicated document archive across versioned S3 storage
- Idempotent pipeline safe to restart or replay at any stage
- Circuit breaker halts scraping after consecutive failures
- Distributed Redis locking prevents duplicate processing across workers
- Exponential backoff with dead-letter queues for all failure paths
- Python
- Playwright
- PostgreSQL
- SQLAlchemy
- Valkey
- AWS
- Docker
- Alembic
- Protocol DI
- 12-Factor App
- Idempotency
- Distributed Locking
Interconnected Terraform Infrastructure
AWS infrastructure for an AI procurement intelligence platform. Four interconnected stacks: shared networking, a queue-driven scraping pipeline, a hardened Flask/Celery web platform, and a self-hosted GPU inference layer, orchestrate everything from VPC to Mistral LLM inference running on GPU EC2 instances inside ECS. A custom Lambda bridge translates Valkey queue depths into CloudWatch metrics to drive ECS autoscaling, bypassing the need for SQS.
Key Impact & Achievements
- Self-hosted 24B LLM on GPU EC2 inside ECS (no SageMaker)
- Lambda queue-bridge: Valkey → CloudWatch → ECS autoscaling
- Four-stack Terraform with explicit remote state dependency chain
- Dual-factor CloudFront origin verification (IP prefix list + secret header)
- Spot/on-demand mixed GPU capacity with rolling deployments
- Workers scale to zero; Lambda-published metrics drive scale-out
- Terraform
- AWS
- Flask
- Celery
- Ollama
- Gotenberg
- Docker
- Python
- Stripe
- Step Scaling
- Service Connect
AWS Infrastructure Scheduler
An intelligent, time zone aware infrastructure orchestration system that automatically manages AWS cloud resources based on business operating hours. Built with Python and leveraging AWS Auto Scaling Groups, ECS, SQS, and Redis, this system reduces cloud compute costs by 50-70% while maintaining 99.9% uptime through sophisticated state preservation and dependency-aware shutdown/startup sequences. 3 years with 20+ semantic versioning releases.
Key Impact & Achievements
- 50-70% reduction in AWS costs
- Zero-downtime operations
- 20+ production releases with backward compatibility
- Multi-timezone awareness and support
- Queue-aware shutdown logic
- 99.9% uptime
- AWS
- Python
- Docker
- Bash
- Kubernetes
- Terraform
- MySQL
- Redis
- CI/CD
- Linux
- State Machine
- Event-Driven
- Microservices
- Idempotency
AWS Infrastructure Library
Python infrastructure library that standardises AWS service interactions across a multi-tenant SaaS platform. Built over 5 years, this internal toolkit abstracts 11 AWS services behind clean, typed interfaces, reducing infrastructure integration complexity by 80% while enabling hybrid cloud deployments. The library serves as the foundational infrastructure layer for a computer vision analytics platform processing millions of video frames daily across hundreds of retail locations.
Key Impact & Achievements
- 15-20% improvement in feature delivery velocity
- Hybrid cloud supporting AWS and Kubernetes
- 11 production-grade service handlers
- 37 production releases over 5 years
- 5-year backward compatibility
- Multi-tenant infrastructure
- AWS
- Python
- CI/CD
- Docker
- Kubernetes
- boto3
- Linux
- Private PyPI
- API Design
- Factory Pattern
- Handler Pattern
- Intelligent Pagination
- Microservices
Intelligent Message Broker
Intelligent message broker for a distributed CV analytics platform. Built with Python, this system autonomously coordinates real-time processing of retail surveillance data across multiple AI engines, handling millions of video frames daily. The platform shows an advanced distributed systems architecture, implementing exactly-once semantics through Redis based state management while maintaining 24/7 autonomous operations.
Key Impact & Achievements
- Cloud-agnostic design
- Multi-tenant scalability
- 24/7 autonomous operation
- Real-time configuration updates
- Thread-safe singleton pattern
- Exactly-once delivery semantics
- Idempotent processing architecture
- GCP
- AWS
- Python
- Kubernetes
- Redis
- MySQL
- Docker
- CI/CD
- Linux
- Singleton Pattern
- 12-Factor App
- Thread Safety
- Idempotency
Synthetic Data Generator
An emergency Python microservice that generates statistically representative synthetic foot traffic data for computer vision analytics systems. Built with exceptional code quality standards (95% test coverage, 115+ automated quality rules), the system maintains data continuity during sensor outages while preserving complex multi-dimensional demographic distributions across temporal patterns.
Key Impact & Achievements
- Sophisticated statistical algorithms with temporal awareness
- Removed need for backup sensor hardware
- Enabled continuous dashboard reporting
- Retroactive backfill capabilities
- Historical pattern analysis
- 99.99%+ Uptime
- Distributed architecture
- Python
- Redis
- MySQL
- Docker
- CI/CD
- Linux
- Data Pipeline
- Microservices
- Test-Driven Development
- State Management
- Algorithm Design
Footfall API
A dual-microservice system that bridges physical retail analytics with cloud-based business intelligence platforms. Built with Python and modern DevSecOps practices, this solution processes customer footfall data from in-store cameras and delivers actionable hourly metrics to external Point of Sale (POS) systems, enabling data-driven operational decisions for multi-location retail chains.
Key Impact & Achievements
- Automated data collection and reporting
- Real-time operational intelligence
- Seamless third-party integration
- Asynchronous architecture
- Batch processing architecture
- 3-month rapid development cycle
- Zero-downtime deployment strategy
- Python
- REST API
- Docker
- CI/CD
- Redis
- MySQL
- Linux
- Flask
- Batch Processing
- Producer-Consumer
- Test-Driven Development
- Stateless Design
Data Persistence Software
An event-driven microservice that serves as the critical data persistence layer for a multi-application computer vision analytics platform. Built with Python, the system processes real-time event streams from AI-powered video analytics and persists structured data into relational databases. Deployed across multiple global regions with 89% test coverage and 4+ years of continuous production operation.
Key Impact & Achievements
- Reduced maintenance overhead by 80%
- Enabled multi-cloud deployment
- Led architectural refactor
- Established code quality culture
- Pioneered modern Python adoption
- 89% test coverage
- Zero-downtime deployments
- AWS
- GCP
- Python
- MySQL
- Docker
- Kubernetes
- CI/CD
- Microservices
- Event-Driven Architecture
- Factory Pattern
- Event Streams
- Real-Time
- Linux
Kubernetes Infrastructure Platform
Architected, built and deployed a cloud-native infrastructure platform that reduced deployment time by 90% (from days to hours). Built using Kubernetes, Helm, and 3,500+ lines of automation, the platform orchestrates 23 infrastructure services and 7 production applications, featuring complete observability (ELK + Prometheus/Grafana), ML inference serving (NVIDIA Triton), and hybrid cloud connectivity. Delivered six-figure annual savings through infrastructure automation and operational excellence.
Key Impact & Achievements
- 90% deployment time reduction
- Six-figure annual savings
- 99%+ deployment success rate
- Zero to production in hours
- 23 infrastructure services orchestrated
- Hybrid cloud architecture
- Complete observability and monitoring
- Kubernetes
- GCP
- AWS
- Helm
- CI/CD
- Terraform
- Docker
- Linux
- Elasticsearch
- Kibana
- Grafana
- Prometheus
- MinIO
- ElasticMQ
- Bash
Digital Preservation Software
A cloud-native surveillance archival platform that automates continuous video recording from commercial security camera systems, reducing storage costs by 60-80% while ensuring compliance with long-term retention requirements. Built with Python, containerised deployment, and multi-cloud architecture (AWS, GCP, on-premises), this system processes video streams 24/7 with microsecond-precision timing, intelligent frame integrity monitoring, and automated credential management.
Key Impact & Achievements
- Continuous 24/7 recording
- Delivered 60-80% reduction in video archival costs
- Self-healing infrastructure maintaining 99.99% uptime
- Agnostic storage supporting AWS, GCP, and MinIO
- 70%+ test coverage with 190+ automated code quality rules
- Created extensible platform supporting 500,000+ cameras
- Containerized deployment using Xvfb (virtual X11 framebuffer) and x11vnc
- Python
- Docker
- AWS
- GCP
- Selenium
- CI/CD
- Microservices
- Compliance
- Linux
- Browser Automation
- Image Processing
- Video Processing
Point of Sale API
Architected and delivered a multi-cloud data integration platform that bridges modern Point of Sale systems with enterprise analytics infrastructure. Built using FastAPI and Python, this webhook-based microservice processes real-time retail transactions across 20+ countries, transforming vendor-specific data formats into standardised analytics schemas. The solution shows expertise in cloud-native architecture, security-first design, and scalable system engineering, delivering sub-second transaction visibility for international retail operations.
Key Impact & Achievements
- Reduced data latency from hours to sub-second processing
- Achieved production deployment in 30 days
- Deployed across 20+ countries with full GDPR compliance
- Automated data pipelines serving 5 business teams
- Engineered vendor-agnostic infrastructure
- Implemented Factory pattern enabling rapid onboarding
- Select-or-insert idempotency patterns
- Comprehensive CI/CD pipeline with pre-commit hooks
- Python
- FastAPI
- Docker
- AWS
- GCP
- REST API
- ETL Pipeline
- CI/CD
- MySQL
- Webhooks
- SOLID Principles
- GDPR
- Real-Time
- Microservices
Multi-Tenant SaaS Cloud Infrastructure
Led the design, implementation, and 5+ year evolution of an Infrastructure as Code platform powering a multi-tenant SaaS application in the retail analytics space. Built using Terraform and AWS, the infrastructure orchestrates 25+ microservices, including ML inference servers, event-driven processing pipelines, and multi-layered data systems. Achieved a 90% reduction in deployment time and 60-70% cost optimisation through intelligent resource management and auto-scaling strategies.
Key Impact & Achievements
- Infrastructure provisioning from 2 days to 2 hours
- 60-70% Total cost optimisation
- 99.99% Production uptime
- Unlimited enterprise scalability
- Modular, reusable architecture
- 30+ production releases with backward compatibility
- 7-Layer Architecture with clear separation of concerns
- GPU-Accelerated ML Infrastructure
- AWS
- Terraform
- Well Architected Framework
- Docker
- ARM
- Linux
- MySQL
- MongoDB
- Microservices
- Event-Driven
- CI/CD
- Blue-Green Deployment
About Me
I'm an entrepreneurial and hands-on technology leader with 11+ years of building profitable, reliable systems from concept to scale. CTO Craft 100 recognised me, alongside CTOs from Netflix, Slack, LinkedIn, and Twitch, for my technical leadership and business impact.
At EVERYANGLE, I led 10+ engineers from founding through €3M+ fundraising to €1M+ ARR with 60%+ profit margins. I excel in building secure and scalable systems that drive measurable business outcomes while maintaining operational excellence, aggressive cost discipline and psychological safety.
Having run the fundraise, managed the P&L, and built the infrastructure that underpinned it, I scope engagements to what actually moves the business — not what's technically interesting.
Currently available for fractional CTO and principal-level infrastructure engagements with funded, resource-constrained companies in the AI and platform space. Typical engagements run 3–12 months — scoped, contracted, and delivered through Wisdom Engineers. Start a scoping conversation
11+ years of professional experience
Software Development, DevOps, Cloud Engineering, and Technical Leadership
6+ years of zero security incidents
Intentional security in depth and system reliability by design
80-90% infrastructure cost reduction
Strategic, step-wise and aggressive infrastructure optimisation
CTO Craft 100 honouree
One of 100 influential technology leaders recognised globally for impact
Remote team leadership
Led distributed engineering teams across multiple time zones — async-capable, high-trust, no co-location required
My Professional Journey
Over 11 years building and scaling production infrastructure. The track record behind the current fractional engagement offer.
Technical Co-Founder
2025 - 2026
Joined an early-stage AI procurement startup as technical co-founder at the pre-MVP stage and built the entire technical foundation across three production systems in approximately five months: an AI/ML document-intelligence and matching service, a multi-tenant B2B SaaS platform, and a data-acquisition pipeline for Irish public procurement notices. Took the product from a non-functional proof of concept to a live platform with tiered subscriptions.
CTO
2019 - 2025
Built the production infrastructure and distributed systems for an AI/ML computer vision retail analytics SaaS platform serving enterprise clients around the world. Led the technical architecture from 0 to 1 while maintaining individual contributor hands on responsibilities. Managed complete AWS/GCP infrastructure supporting 25+ microservices processing millions of video analytics events daily across thousands of retail locations. The company raised €3M+ in seed funding.
Firefighter
2024 - Present
Serving the community as a retained firefighter responding to emergency calls. Demonstrating commitment to community service involves crisis and risk management, making rapid decisions under pressure, and coordinating teams in high stakes environments.
- Respond to fire, search, rescue, and emergency medical incidents across County Cavan in Ireland as part of a highly trained on-call retained firefighter
- Completed intensive firefighter training including breathing apparatus, compartment fire behaviour, casualty rescue techniques, hazardous materials response, and emergency medical care
- Maintain emergency response readiness with regular training drills, equipment maintenance, and physical fitness
Career Transition
2017 - 2019
Strategically pivoted from web consulting to formal software engineering through intensive formal education and hands-on project work, recognising the shift toward cloud-native infrastructure and enterprise systems. I used this period for intensive technical upskilling, emergency services training, and volunteer leadership.
Founder & CEO
2012 - 2017
Founded and operated a profitable full-service technical consultancy company delivering custom software, cloud solutions, DevOps, and IT infrastructure to international enterprise clients. Kept profit margins between 50% and 60% for five years in a row. Direct proof of sustainable independent technical delivery at commercial margins.
- Architected and deployed cloud-hosted web applications and SaaS platforms using AWS EC2, RDS (MySQL/PostgreSQL), S3, CloudFront CDN with automated backup strategies and monitoring via CloudWatch
- Implemented CI/CD pipelines for client deployments using Git-based workflows, automated testing, and blue-green deployment patterns, minimising downtime
- Provided 24/7 infrastructure support and incident response with a 3-hour SLA for critical issues, maintaining 99% uptime for critical business systems
- Designed scalable backend systems (PHP, MySQL, JavaScript) with payment gateway integrations (Stripe, PayPal), third-party API connections, and responsive frontend
- Delivered 50+ technical projects including e-commerce platforms, booking systems, custom CRM/CMS solutions, and corporate intranets with 80% client retention and 60% referral-based revenue
Let's Connect
If you're scaling infrastructure, building AI/ML platforms, or need senior technical leadership without a full-time hire let's talk.
My Email Address
My LinkedIn Profile
- Remote-first, Irish/UK/EU time zones
- Fractional CTO or Principal Engineer
- Platform builds, AI/ML infrastructure, and agentic tooling
- Seed through Series C startups
- Day-rate or retainer, scoped via Wisdom Engineers
- Async-first, distributed teams