Freelance Webhook Platform H/F - Collective Works

The product you'll work on

Roundtable is the infrastructure powering private market investments. Investors, asset managers and other private market professionals use it to manage deals, commitments, investments and the workflows around them.

Our Public API already lets customers read their data. What it can't do yet is tell their systems when something changes, so today they have to poll. Our customers want to push Roundtable data into their own systems or automation tools the moment a deal is updated or a commitment is signed.

We're building a webhook platform so external applications can subscribe to events across the investment lifecycle. The foundations are there: an internal event bus whose events are published transactionally through a pg-boss outbox, background jobs on PostgreSQL, a Public API and a first design for the data model and delivery flow.

We're looking for a senior full-stack engineer who has already built webhook infrastructure, not just consumed third-party webhooks.

What you'll build

Here are the challenges on our list:

Event catalog. Build on our event bus to define consistent business events, with payload schemas, naming and versioning built in from day one. Payloads follow our Public API conventions.

Reliable delivery. Build the webhook consumer on top of our event outbox and deliver asynchronously with retries, exponential backoff and dead letters. Decide the delivery guarantees, the retry policy, and when a failing endpoint gets disabled.

Security. Signed payloads with replay protection, secret rotation without downtime, endpoint validation against SSRF, and strict tenant isolation: an organisation only ever receives events about its own data.

Observability. Trace every event from emission to each delivery attempt, wired into Datadog and OpenTelemetry, so we see failures before customers do.

Developer experience. Build the screens to manage subscriptions, send test events, inspect delivery history and replay failures, plus the documentation an integrator needs.

How you'll work

This is a hands-on engineering role. You'll work with our CTO and engineers to design and ship the solution. Depending on the mission, spending some time with the team at our Paris or Toulouse offices can help, and we'll agree on it together.

We expect you to:

• Own the webhook platform end to end, from architecture to production

• Build on our existing event bus and job infrastructure rather than reinventing it

• Challenge our assumptions and make pragmatic technical decisions

• Balance reliability and security with the need to ship quickly

• Build monitoring and debugging tooling from day one

• Ship incrementally, validate with real customer use cases and iterate

We're not looking for weeks of architecture diagrams. We want sound decisions, production-quality code and working infrastructure.

The stack that matters day-to-day

• Language: strict TypeScript, no any

• Frontend: Next.js App Router, React Server Components, Tailwind CSS, shadcn/ui, Radix UI

• Backend: Node.js, tRPC, Zod, PostgreSQL, Prisma, OpenAPI

• Async processing: internal event bus, pg-boss, Redis, BullMQ

• Infrastructure: GCP, Turborepo

• Observability: Datadog, OpenTelemetry

Mission details

• Contract: freelance

• Location: Paris or Toulouse

• Schedule: full-time

• Duration: a defined first scope, which can be extended

• Start date: as soon as possible

• Daily rate: to be discussed

What we're looking for

Real webhook infrastructure experience. This is the most important requirement. You've designed and run a webhook delivery system in production, ideally in a SaaS or API platform. You've faced several of these:

• Event catalogs, payload schemas, versioning and subscription models

• Async delivery pipelines with retries, dead letters and replay

• Duplicates, ordering, idempotency and delivery guarantees

• Signatures, replay protection and secret rotation

• Delivery logs, monitoring and debugging tools for customers

You can explain the tradeoffs behind your past implementations, including what went wrong and what you would do differently today.

An AI-native workflow. You use Claude Code, Codex or similar coding agents every day, and you know how to check AI-generated code on critical infrastructure. Ownership stays human.

Domain Curiosity. You don't need to know private markets yet. We'll teach you the domain.

• API
• Intelligence artificielle
• PostgreSQL
• Node.js
• Google cloud platform
• Isolation
• TypeScript