Join us
Retab is an applied research lab building the automation platform of the future: A one-stop-shop where companies can seamlessly experiment and integrate AI into their most critical workflows.
We believe deploying AI at scale isn't a one-time project, but a living process of iteration, learning, and mastery. retab gives teams the power to move at the speed of their ideas, building smarter, faster, and more reliable systems that drive lasting business transformation.
We are backed by VF, K5 Global (Uber, SpaceX, xAI), Eric Schmidt's European investment fund, and the founders of Datadog, Dataiku and Free.
We are a group of researchers, engineers, and builders. We build software to invent at the edge of what's possible. Our work has already improved the lives of hundreds of companies.
If this excites you, we'd love to hear from you.
Open Positions
Our stack: React (Nextjs), Python (FastAPI). No matter the role, you will operate autonomously on the whole codebase. We grind hard and believe that great work happens within small, tight-knit teams.
Full-Stack Engineer — Product+
Some of the largest companies in the US and Europe rely on Retab to automate document-heavy workflows with AI. As the product grows, we need engineers who can own features end-to-end — from database schema to pixel-perfect UI.
We're looking for a strong full-stack engineer to design and ship product features that make Retab delightful and powerful for our users.
We believe that design is a powerful form of engineering — and that true frontend beauty is deeply rooted in the thoughtful design of the backend.
Example problems you might work on
- • How should a real-time collaboration experience work when multiple users are editing the same extraction schema?
- • A customer needs to review hundreds of extracted documents. How should batch review interfaces be designed so they stay fast and ergonomic?
- • How should we design workflow builders that feel intuitive to non-technical users while exposing the full power of the platform?
- • A new API endpoint needs to handle large file uploads with progress tracking, retries, and graceful error handling across the stack.
What you'll build
- • Core product features — extraction editors, document viewers, workflow UIs
- • API design — clean, well-documented endpoints that power both the dashboard and the SDK
- • Performance improvements — making complex pages fast with virtualization, caching, and smart data loading
- • Developer experience — improving internal tooling, testing, and CI workflows
Stack: Next.js, React, Python (FastAPI), MongoDB.
Our inspiration comes from industry leaders like OpenAI, Stripe, and WorkOS, whose relentless pursuit of excellence guides our own ambitions.
You don't need experience with every part of the stack, but you should be comfortable picking up new pieces quickly. We also expect engineers to be comfortable with modern coding tools like Codex and Claude Code.
We care deeply about product quality. The best interfaces come from engineers who understand that great UX starts with well-designed systems.
Infrastructure Engineer — Scaling LLM Systems+
Retab processes millions of documents for some of the largest companies in the US and Europe. Behind the product is a distributed system where a single request might trigger OCR, embedding, and multiple LLM calls across a fleet of specialized services — and any of those steps can take minutes, fail silently, or spike memory by an order of magnitude.
This is not a typical SaaS backend. Workloads are bursty and unpredictable. A request could be a single invoice or a 300-page legal contract. LLM providers degrade without warning. Workers need to scale across dimensions that don't fit neatly into CPU/memory autoscaling. The infrastructure problems here are genuinely hard.
We need an engineer who can design and operationalize these systems at a high level — someone self-driven who can own the reliability and scalability of the platform with minimal hand-holding.
The kinds of problems you'll face
- • An LLM request has been running for 10 minutes. Is the provider degraded, or is this a legitimately large document? How should the system tell the difference?
- • A workflow keeps retrying because an upstream API is partially down. How should retries, backoff, and circuit breaking behave to avoid cascading failures across the fleet?
- • A customer submits 50,000 documents at once. How should the system prioritize, queue, and shed load without starving other tenants?
- • MongoDB is under heavy write pressure from async workers. What should move to queues, caches, or event-driven patterns?
- • The Go orchestrator and Python services need to coordinate long-running workflows. How should health checks, timeouts, and observability work across language boundaries?
What you'll own
- • Infrastructure-as-code — our entire GCP footprint is managed with Terraform (Cloud Run, Memorystore, Cloud SQL, GCS, networking)
- • Service orchestration — scaling and reliability of our Go orchestrator, Python API gateway, and specialized worker fleets
- • Data infrastructure — MongoDB performance, Valkey caching layers, storage patterns for large artifacts
- • Observability — Grafana dashboards, alerting, distributed tracing across a polyglot system
- • CI/CD and deploy pipelines — making deploys fast, safe, and boring
Stack: GCP, Temporal, MongoDB, Valkey, Sentry, Terraform, Grafana.
You should be proficient in at least Python or Go, comfortable with Terraform, and familiar with distributed systems. We also expect engineers to be comfortable with modern coding tools like Codex and Claude Code.
You should be an expert in your field. We care about strong systems thinking and the ability to operate independently — not about checking every box on a technology list.
Full-Stack Engineer — Enterprise Ready Features+
Some of the largest companies in the US and Europe rely on Retab to automate document-heavy workflows with AI. As adoption grows, we're building the enterprise layer of the product.
We're looking for a strong full-stack engineer to design and ship the systems companies rely on to adopt Retab at scale.
Enterprise features often look simple in the UI but hide a lot of complexity underneath: identity providers, permission systems, billing models, and admin tooling. Our goal is to make these systems powerful but intuitive.
Example problems you might work on
- • A company connects their identity provider via SAML SSO. How should users be provisioned, mapped to organizations, and handled when identities change?
- • How should a role-based access control system be designed so permissions stay understandable as the product grows?
- • A customer's usage suddenly increases. How should billing and usage tracking behave so it stays transparent and predictable?
- • How should enterprise admin interfaces expose powerful configuration without overwhelming users?
What you'll build
- • SSO integrations (SAML / OAuth)
- • Role-based access control (RBAC)
- • Billing and usage systems
- • Enterprise admin and settings interfaces — Building a great product experience is key!
Stack: Next.js, Python, Stripe, WorkOS.
You don't need experience with every part of the stack, but you should be comfortable picking up new pieces quickly. We also expect engineers to be comfortable with modern coding tools like Codex and Claude Code.
We care deeply about product quality. The best interfaces come from engineers who understand that great UX starts with well-designed systems.
Developer Advocate (Developer Relations)+
As our first Developer Advocate, you'll sit at the intersection of engineering, storytelling, and community. Your mission: help developers discover what's possible with Retab—and make their experience seamless and exciting.
What You'll Do
- • Create world-class content: Write tutorials, record demos, and publish examples that make developers want to build with Retab.
- • Build sample apps: Showcase integrations with Python, TypeScript, and popular frameworks.
- • Engage the community: Be active on Slack, Discord, and social channels—answering questions, amplifying projects, and nurturing trust.
- • Represent Retab publicly: Speak at events, join podcasts, and produce content that tells our story with clarity and authenticity.
- • Advocate for developers: Gather feedback, influence product direction, and improve docs and onboarding.
Who You Are
- • You're both a developer and a teacher—excited to explain complex ideas simply.
- • You have hands-on experience building with Python or JavaScript/TypeScript, and ideally with APIs, AI, or data processing tools.
- • You've published technical content (blog, video, talks) that's genuinely useful.
- • You're empathetic, approachable, and comfortable representing a product in public.
- • You thrive in fast-moving, small-team environments where you can own your work end-to-end.
Nice-to-Haves
- • Prior DevRel or community experience
- • Familiarity with AI/ML or document automation
- • Existing online presence (blog, GitHub, X, YouTube, etc.)
- • Experience with open source or developer tooling
Why Join Retab
- • Massive impact: You'll shape Retab's entire developer story and community.
- • Cutting-edge tech: Work on real AI problems that power how companies process documents.
- • Ownership: Join a small, well-funded team where your ideas ship fast.
- • Flexibility: Remote-friendly with hubs in Paris and San Francisco.
If you've created content (blogs, videos, or demos), we'd love to see it.
Forward Deployed Agent Engineer+
Large organizations across the US and Europe rely on Retab to process and reason over massive volumes of documents. To make this possible, we built a powerful Model Context Protocol (MCP) server that exposes Retab's capabilities to AI agents and developer tools — a forward deployed engineer in software form.
But real-world systems are messy. Automation rarely solves everything on its own. We are looking for Forward Deployed Agent Engineers who can bridge that gap — sitting at the intersection of engineering, AI systems, and real-world customer deployments.
You will work directly with companies to design, build, and deploy automation systems on top of Retab, while continuously improving the MCP itself so that more of this work becomes automated over time.
What you'll do
- • Ensure the MCP works in the real world — maintain, extend, and stress-test the MCP by deploying it in real environments. Work with the core engineering team to improve the protocol, tools, and developer experience.
- • Work with customers end-to-end — understand complex operational workflows, design document processing pipelines, integrate Retab with internal tools, and build prototypes and production-grade systems.
- • Solve the last mile — handle the final 10–20% of problems that turn a promising prototype into a system a company can rely on every day: custom integrations, edge cases in document pipelines, new workflows, and debugging complex automation systems.
Example problems you might work on
- • A company wants to automate a multi-step workflow involving document ingestion, classification, extraction, validation, and downstream system updates.
- • A customer processes hundreds of thousands of documents per week and needs a reliable pipeline.
- • Designing an AI-agent workflow using the MCP to orchestrate multiple tools and document-processing steps.
- • Extending the MCP so that workflows that previously required custom engineering become fully automated.
Who thrives in this role
- • You enjoy solving ambiguous real-world problems and move quickly from idea to working system.
- • You care deeply about product quality and like working directly with users.
- • You take ownership of problems end-to-end — from debugging a production pipeline in the morning to designing a new system in the afternoon.
Nice-to-Haves
- • Experience as a Forward Deployed Engineer, Solutions Engineer, or Founding Engineer
- • Experience building automation or workflow systems
- • Familiarity with AI agents or LLM-based systems
- • Experience with developer platforms or infrastructure
Stack: Python, Next.js, AI / LLM infrastructure, Stripe, WorkOS.
You don't need experience with every part of the stack. We care more about engineering ability and ownership.
Interview process
If your application clears the initial review, you will be invited to a 30-minute phone interview where a team member will ask you some basic questions. If you pass this phone interview, you will proceed to the main selection process, which includes:
- • A coding interview (1h): You guide us through some samples of a piece of code you are proud of and explain us how you solved the problems you encountered.
- • A team interview, where you will meet more team members to see if you fit our culture
We aim to complete this process within one week. All applications are evaluated by our technical team members.
Please email your resume and a note on a project you're proud of to careers@retab.com.