Implementing a Reliable Workflow Engine in TypeScript

How to implement a reliable workflow engine in TypeScript - SFNode presentation

I had a great time presenting DBOS Transact at the SFNode Meetup in San Francisco on May 31, 2024. My presentation focused on why reliable workflows make many business problems easier to solve (for example when doing payments or reliable event processing), why they're hard to implement right now, and how the DBOS Transact open source durable execution library (Python and TypeScript) makes workflows easy to use.

Here are the slides I presented

There were some good questions from the audience:


Q: Why did we choose to implement DBOS Transact in TypeScript?
A: TypeScript's support for first-class functions and robust typing means we can implement workflows with durable execution at the library level, so developers can get their benefits without radically changing the way they code.

Q: How does DBOS handle application updates while workflows are actively in process?
A: Each workflow is versioned, so new requests go to servers hosting new application versions while old workflows can be processed in the background by a different server hosting an old version.

If you have questions about the slides or DBOS, feel free to ask them on the DBOS Community on Discord. I look forward to hearing from you.

Insights

Recent articles

The latest in durable execution, AI workflows & more.

Benchmarks
Apr 23, 2026

Does Postgres Scale?

Benchmarking the workflow execution and workflow queueing scalability of a single Postgres server.
Peter Kraft
How To
Apr 21, 2026

Making AI Agents Stateful and Fault-tolerant on Google Cloud Run

Using open source DBOS durable execution libraries to make Google Cloud Run a safe place to run failure-proof, stateful agentic AI workflows.
Henri Maxime Demoulin
Product news
Apr 13, 2026

DBOS Enhancements - April 2026

Overview of new DBOS durable workflow execution and workflow ops features, including cross-language workflow interoperability, a new metadata-only mode for increased data privacy, LlamaIndex and Databricks LakeBase integrations, and more.
Qian Li