Magic (makers of Loyalist) · New York, NY · In-office 4 days/week · $170k-$210k + equity
I’m Chris, a founding engineer at Magic. You’d be taking over a lot of the infrastructure work I’ve been holding alone, so I wanted to tell you about it from my perspective.
I came to Magic from Hightouch, a martech startup where I worked on data pipelines and syncing for companies across industries. I wanted to go somewhere smaller, more in person, and plugged into a real-world, NY-centric industry. At Magic, we’re supporting a legacy hospitality industry just starting to adopt serious tech, and we have customers right outside the office in Soho.
Year one got us to 800 restaurants. Year two is about building for 1,000 and beyond and growing to new verticals. You’d be the one architecting that.
Loyalist is the operating system behind how 800+ restaurants remember their guests. We ingest live data from ~30 third-party systems before, during, and after service (reservations, POS, reviews, private events, email, SMS, e-commerce, data warehouses) and turn it into action. We track 40 million guest profiles across our customers, hundreds of millions of historical visits and orders, and more than 10 million personalized guest interactions every year.
Customers run the gamut, from favorites literally in my neighborhood like Place des Fêtes to Carbone, Momofuku, Le Bernardin, and Daniel Boulud’s group. The largest of them have 5-10 million guests each and import 15 years of history in one go. (Some of them are precise enough about their data to email us when they notice a specific guest’s 2010 visit didn’t import.)
Under the hood, Loyalist is closer to a real-time data platform than a CRM. Tons of background jobs importing data, syncing guest profiles and tags, and sending email and SMS automations. Those jobs are set-and-forget on their side and mission-critical on ours. That asymmetry shapes our infra work.
A few of the problems I find most interesting:
Silent pipeline failures we have to catch before the restaurant does. If a Toast instance drops on a Wednesday, data stops flowing into Loyalist. The customer won’t notice (they’re running their restaurant). We might not notice for a few days either, and by Saturday the revenue report is wrong and we’ve missed three days of customer journeys. We’ve made a lot of progress here but still don’t have enough of the right signal in Datadog. This is one of the first things you’d work on.
Background job orchestration at scale. We’re running on Celery, which has carried us to 800 restaurants, but it’s probably not the right tool for where we’re going. Migrating to something modern like Temporal is a project a senior engineer would scope and own.
Analytical workloads. Some customers import 15 years of history and then want to run reports against it. Right now those queries run on the same database that’s syncing live reservations during dinner service! Moving analytical workloads to something purpose-built is something I want you weighing in on from day one.
You’ll own our AWS infrastructure: RDS, background jobs, caching, networking, etc. You’ll also own a lot of the integration layer, like a new POS integration, sometimes talking with partner companies.
First 30–60 days, I’d expect you to focus on those three things: getting our observability to actually tell us what’s broken and how badly, scoping the background job orchestrator migration, and scoping the analytical database migration. Ideally we make meaningful progress on these fast.
You’ll talk to customers, usually when they’re asking for a new integration or surfacing a specific issue and we need someone who can speak to the infra implications on a call. You’ll also partner closely with Niky and the product side.