Asks for

The ad says“Real comfort at the systems level: async runtimes, backpressure, memory layout, lock contention, what a query planner is actually doing”
The ad says“Real comfort at the systems level: async runtimes, backpressure, memory layout, lock contention, what a query planner is actually doing”
The ad says“Rust for most of the engine — we want a long-running service that holds a lot of state and doesn't surprise us on a Tuesday. Go for ingress, where the job is mostly IO and protocol translation. Protobuf and gRPC between services, with the proto schema as the contract. Python for the SDK and IaC tooling. Postgres and Redis for state, Kubernetes and Helm to deploy, GitHub Actions for the rest. Claude, Gemini, and Codex in daily use — we ship weekly, and we don't pretend AI didn't write what it wrote. You don't need to have used every warehouse on the list; you do need to be the kind of person who reads the dialect's docs when a ticket lands on it.”
The ad says“Rust or Go shipped and maintained in production — a service that paged you at 3am, and that you then made stop paging you”
The ad says“Rust for most of the engine — we want a long-running service that holds a lot of state and doesn't surprise us on a Tuesday. Go for ingress, where the job is mostly IO and protocol translation. Protobuf and gRPC between services, with the proto schema as the contract. Python for the SDK and IaC tooling. Postgres and Redis for state, Kubernetes and Helm to deploy, GitHub Actions for the rest. Claude, Gemini, and Codex in daily use — we ship weekly, and we don't pretend AI didn't write what it wrote. You don't need to have used every warehouse on the list; you do need to be the kind of person who reads the dialect's docs when a ticket lands on it.”
The ad says“Real comfort at the systems level: async runtimes, backpressure, memory layout, lock contention, what a query planner is actually doing”
The ad says“Real comfort at the systems level: async runtimes, backpressure, memory layout, lock contention, what a query planner is actually doing”
The ad says“Rust for most of the engine — we want a long-running service that holds a lot of state and doesn't surprise us on a Tuesday. Go for ingress, where the job is mostly IO and protocol translation. Protobuf and gRPC between services, with the proto schema as the contract. Python for the SDK and IaC tooling. Postgres and Redis for state, Kubernetes and Helm to deploy, GitHub Actions for the rest. Claude, Gemini, and Codex in daily use — we ship weekly, and we don't pretend AI didn't write what it wrote. You don't need to have used every warehouse on the list; you do need to be the kind of person who reads the dialect's docs when a ticket lands on it.”
The ad says“We read from twenty-something source systems, run validators against the data at real volume, and raise incidents the moment something is wrong. You'll own pieces of it end to end: the protobuf contract, the Rust service behind it, the SQL it generates, the worker loop that runs it, and the metrics that say it's healthy.”
The ad says“Rust for most of the engine — we want a long-running service that holds a lot of state and doesn't surprise us on a Tuesday. Go for ingress, where the job is mostly IO and protocol translation. Protobuf and gRPC between services, with the proto schema as the contract. Python for the SDK and IaC tooling. Postgres and Redis for state, Kubernetes and Helm to deploy, GitHub Actions for the rest. Claude, Gemini, and Codex in daily use — we ship weekly, and we don't pretend AI didn't write what it wrote. You don't need to have used every warehouse on the list; you do need to be the kind of person who reads the dialect's docs when a ticket lands on it.”
The ad says“Real comfort at the systems level: async runtimes, backpressure, memory layout, lock contention, what a query planner is actually doing”
The ad says“Rust for most of the engine — we want a long-running service that holds a lot of state and doesn't surprise us on a Tuesday. Go for ingress, where the job is mostly IO and protocol translation. Protobuf and gRPC between services, with the proto schema as the contract. Python for the SDK and IaC tooling. Postgres and Redis for state, Kubernetes and Helm to deploy, GitHub Actions for the rest. Claude, Gemini, and Codex in daily use — we ship weekly, and we don't pretend AI didn't write what it wrote. You don't need to have used every warehouse on the list; you do need to be the kind of person who reads the dialect's docs when a ticket lands on it.”
The ad says“Rust or Go shipped and maintained in production — a service that paged you at 3am, and that you then made stop paging you”
The ad says“SQL well beyond SELECT — window functions, query plans, dialect differences, and debugging by reading EXPLAIN rather than guessing”

Validio is the agentic data management platform that automates data quality, lineage, and cataloging — so enterprises can finally trust the data behind their analytics, AI, and reporting. As a Backend Engineer on our Validation Engine team, you build the part that does the actual work — the engine customers pay us for.

We read from twenty-something source systems, run validators against the data at real volume, and raise incidents the moment something is wrong. You'll own pieces of it end to end: the protobuf contract, the Rust service behind it, the SQL it generates, the worker loop that runs it, and the metrics that say it's healthy.

What you'll do

Own parts of the engine end to end, from proto contract to production metrics
Build source connectors for the warehouses customers ask for — schema introspection, query generation, auth, the full path
Generate correct, performant SQL across every dialect our customers run on
Extend our lineage parser into the dialect corners it doesn't cover yet
Debug hard systems problems — dropped watermarks under load, lock contention, query plans that don't scale — and leave the regression test behind
Work directly with Customer Success, Solutions Engineering, and Support, and be on the call when the customer describes the problem

What you'll bring

Rust or Go shipped and maintained in production — a service that paged you at 3am, and that you then made stop paging you
Real comfort at the systems level: async runtimes, backpressure, memory layout, lock contention, what a query planner is actually doing
SQL well beyond SELECT — window functions, query plans, dialect differences, and debugging by reading EXPLAIN rather than guessing
An instinct for invariants: you can say what a change preserves, what it breaks, and how it fails before it ships
The judgment to direct AI tools well — when to lean on them, what they get subtly wrong, how to keep a codebase coherent when half the diffs aren't yours

Bonus points for

Experience with data quality, observability, lineage, or catalog systems
Air-gapped or on-prem deployment, and Kubernetes and Helm at production scale

Our stack, and why

Rust for most of the engine — we want a long-running service that holds a lot of state and doesn't surprise us on a Tuesday. Go for ingress, where the job is mostly IO and protocol translation. Protobuf and gRPC between services, with the proto schema as the contract. Python for the SDK and IaC tooling. Postgres and Redis for state, Kubernetes and Helm to deploy, GitHub Actions for the rest. Claude, Gemini, and Codex in daily use — we ship weekly, and we don't pretend AI didn't write what it wrote. You don't need to have used every warehouse on the list; you do need to be the kind of person who reads the dialect's docs when a ticket lands on it.

Why Validio

Founded in Sweden in 2019, we're trusted by data-driven enterprises like Nordea, Canva, Truecaller, Traveloka, and Deutsche Glasfaser. We raised a $30M Series A this year and grew ARR 800% in twelve months as enterprises got serious about making their data AI-ready. At around 25 people, you'll own initiatives end to end and make the call when you're closest to the problem. We're strict about testing and documentation, and fast about decisions.

Come as you are. We hire for what you'll build with us, and we welcome applicants of every background.

Asks for

Values