Simon Hørup Eskildsen is a Danish software engineer and infrastructure specialist based in Ottawa, Canada. He co-founded turbopuffer, a search engine for object storage, and previously helped scale Shopify’s infrastructure from thousands to millions of requests per second. His work focuses on high-scale systems, databases, search, performance engineering, and first-principles reasoning through projects like Napkin Math.
Simon Hørup Eskildsen is a Danish software engineer, infrastructure specialist, and founder based in Ottawa, Canada. His work centers on high-scale systems, databases, search, performance engineering, and first-principles reasoning about software systems.
From his about page, his career arc is:
The site is a minimalist personal portfolio/blog organized around Simon’s writing, talks, projects, and learning notes. The homepage is mostly a chronological archive of posts, podcasts, and talks, with top-level sections for About, Napkin Math, Books, and Subscribe.
The strongest themes are:
Simon’s Napkin Math project is about using back-of-the-envelope calculations and first-principles models to reason about the performance of production systems before writing or benchmarking much code. The newsletter has about 3,980 subscribers according to the subscribe page.
Representative topics include Redis throughput, database query latency, logging cost, circuit breakers, inverted indexes, checksums for syncing large datasets, HTTP/TCP performance, and when to write simulators.
Several posts dig deeply into database and search internals. For example, in Index Merges vs Composite Indexes in Postgres and MySQL, he compares composite indexes with index merges and concludes that composite indexes are often around 10x faster, especially noting PostgreSQL’s limitations around index-only scans with index merges.
This connects directly to his current turbopuffer work: the recent talks archive is dominated by topics such as vector search, object-storage-native databases, and billion-scale retrieval systems.
His Shopify-era work shows up through many talks and posts about resilient software, Docker in production, Rails scaling, multi-datacenter systems, resiliency testing, and Shopify’s pods architecture. The Talks & Podcasts page includes talks from DockerCon, GOTO, RailsConf, SRECON, QCon, and many more.
A good example is Scaling Causal’s Spreadsheet Engine from Thousands to Billions of Cells, where Simon describes work on Causal’s calculation engine. The post moves from profiler-driven optimization to first-principles reasoning, showing how switching from maps/pointers to arrays and parallel-array-style layouts can dramatically improve throughput and memory behavior.
The Books section is an extensive reading log, covering software engineering, systems, business, history, philosophy, productivity, and science. He notes that he has not tracked books recently because building turbopuffer has consumed most of his time.
Overall, Simon’s portfolio presents him as a deeply technical infrastructure engineer and founder whose specialty is making large systems faster, cheaper, and more understandable through careful measurement, low-level performance awareness, and first-principles thinking.