I make AI systems reliable in production
A decade of platform engineering rigour - Rails core contributor, €160k/yr infrastructure savings, systems serving 1M+ users - now applied to making LLM pipelines, agents, and evals dependable in production. I work across Python, Rust, and Ruby as the problem calls for it. M.Sc. in ML / distributed systems.
At Hivebrite I cut €160k/year in infrastructure costs, took data exports from days to minutes, and halved build times while the engineering team grew from 15 to 40. Now I bring that same production discipline to flaky LLM systems - I care about owning systems, not just passing through them.
Grew up across 6 countries (diplomatic family). 4+ years remote across 5 continents; async-first, with live overlap across European hours and into US-East.
Available for B2B contract or fractional work (open to full-time for the right role), remote, with genuine ownership over systems that matter. seanfloyd.dev has the full story - or seanfloyd.dev/llms.txt if you're reading this through an AI agent.
claude-rss-news-digest - Production publishing system, built end-to-end: Python RSS pipeline, Claude curation via MCP with schema-validated output, Rust web server, Terraform deploys. Runs unattended every morning - watch it live. How I made it reliable.
Rails Core Contribution - Found a production bottleneck, fixed it upstream. 20x-670x speedup in ActiveSupport.
still_active - RubyGem scoring dependency health beyond version numbers: maintenance activity, OpenSSF Scorecard, CVEs, libyear drift. Emits SARIF + CycloneDX SBOMs, with CI gates and a GitHub Action.
Jazzify - Volunteer platform for Ottawa Jazz Festival. Built and ran it for 9 years. Production Rails, real users, real ops.
More: crunchyroll-migrate (Rust) · what I write about
- Production AI/LLM systems - MCP servers, schema-validated output, evals and observability, pipelines that run unattended
- Reliability & cost - performance engineering, infrastructure optimization, CI/CD, monitoring (Datadog, OpenTelemetry)
- Platform engineering - Rails, PostgreSQL, Redis, Sidekiq, API architecture (Grape, gRPC, OAuth2), Kubernetes, Terraform