How we used triplet loss and embeddings to improve accounting on Ramp.
How Ramp applied Thompson Sampling to improve bank linking success rates
How we modernized a machine learning codebase to increase velocity, quality, and confidence
How we applied engineering and product principles to scale Ramp to become one of the fastest growing software businesses of all time
Reflections on creating and scaling the Core Engineering team
The genesis of Ramp’s workflows product and lessons learned
How Ramp achieved just-in-time access in AWS
With the passage of time, small inefficiencies can add up to big bottlenecks
A walkthrough of why and how Ramp uses Metaflow for machine learning engineering
The lessons I learned tech leading bill pay from 0 to 1
A deep dive into why investing in scalability, reliability, and extensibility from Day 1 embodied our mantra of velocity at all costs.
Principles for developing AI models that consistently deliver meaningful outcomes
Questioning what is gained from "clean code" in actual, production environments.
How Ramp tackled runaway Snowflake costs.
How we built a general-use Redis-backed rate limiter
Unblocking engineers by letting them merge their code 5x faster.
A couple quick wins and a bigger bet to speed up our CI test durations
Building — and adopting — a component library in the context of a vibrant business is no easy feat. Here are a few things we've learned.
The Ramp internship experience, from a former intern
How we built Ramp by taking asymmetric risks, and why you should, too
Seven principles Ramp uses to develop high-quality products, quickly
Three months of learning and building at New York City's fastest-growing startup.
Pablo talks about how and why we use Elixir at Ramp.