Teams building AI applications need to monitor and evaluate them continuously to get them into production and keep improving them. Langfuse is the open-source platform for that workflow. Read about our mission.

In January 2026, we joined ClickHouse. We still run as our own team.

If you want to build great developer tools, or get developers to love them, we'd love to hear from you.

– Marc, Clemens, Max and the Langfuse team

Why join Langfuse

We're winning the category through open source

Langfuse is the most used open-source AI engineering platform, with the largest distribution in our space. See the public metrics.

Langfuse is the most widely adopted LLM Engineering platform, used by 50,000+ companies, with 37,123 GitHub stars, 65M+ SDK installs per month, and 38M+ Docker pulls. Trusted by 21 of the Fortune 50 and 129 of the Fortune 500 companies.

AI-native companies and large enterprises

AI-native companies like Ramp run on Langfuse, and so do very large enterprises. See more companies building with Langfuse.

You own an area, and your work is public

Engineers own a feature area and stay with it as the team grows. Marketers own launches, events, or a program like Langfuse Academy. The people who build a feature answer users in GitHub Discussions themselves. Because the work is open source, your track record stays public.

You own it from the customer call to the launch. Here is how that looked for one real feature, the evals revamp:

1. Customer. It starts with users telling us in GitHub Discussions that mapping data into evaluators was painful. (How we work with customers)
2. RFC. You write the RFC. The team reviews it async, then we settle the hard questions in one intense session. (Roadmapping)
3. Build. Then you build. A lot. The main PR touched 600 files. (How we ship)
4. Ship. You launch it with the GTM team: changelog, docs and marketing. (Releases)
5. Offsite. Then you meet everyone in person at an offsite in Tokyo or Berlin. (Remote at Langfuse)

ClickHouse backs us and stays out of the way

We set our own roadmap. ClickHouse gives us a global sales team that sells Langfuse and a much larger audience for our launches. The engineers who build the database are one message away.

Our bet is that ClickHouse will be one of the great infrastructure companies of the AI era, and that joining Langfuse now means being part of that. The database already runs real-time analytics, observability, and AI workloads. Langfuse is how that platform sees what agents do in production. For more background, see below.

AI-pilled and remote

Everyone gets a very high AI budget. We build slides with AI, maintain our own coding harness, and almost everyone runs a small army of agents. We have three scheduled meetings a week. Everything else happens in writing (see our principles).

What you'll work on

Each of these projects was usually executed by one person.

Product engineering

You own your projects end to end and the product outcome, building whatever it takes, putting the user and the product first, and making smart technical choices.

Data platform

You own ingestion, ClickHouse query performance, and the infrastructure behind Langfuse Cloud and self-hosting, at petabyte scale.

Marketing and growth

Our audience is AI engineers, so most of our marketing is material they'd bookmark:

• Developer content, like the changelog, Langfuse Academy, and the docs
• Launches with one owner each, which also go out through ClickHouse's much larger audience
• Product marketing, events and community, and growth

Open roles

We are remote-first and hire engineers across Europe and in US-East time zones. Engineering teams are mixed EU/US. Marketing and all other roles hire in Europe and all of the United States. All roles are listed on the ClickHouse careers site. See how we think about pay and perks.

Team

Further reading

ClickHouse

Videos and podcasts

Handbook

We document our principles and processes in public, so you can decide whether you'd enjoy working here.

Open source

Almost everything we do is public:

Public metrics

---

---