Data Engineering Conferences 2027

"Data Engineering Summit" is the proper name of 1 event: the conference AIM Media House runs in Bengaluru, whose 6th edition is on 13 to 14 May 2027. Most people who search the phrase mean a looser thing, a summit where data engineering is the main subject, and the 4 vendor events in the list cover that ground for their own platforms.

The word "data engineering" also names a research field. The IEEE International Conference on Data Engineering, ICDE, is that field's flagship, and its 2027 edition is the 43rd, in Copenhagen on 17 to 21 May 2027. Its sessions are about query processing and storage in data management systems, so a practitioner looking for pipeline war stories will find few. The 3 research conferences in the list are worth following for a different reason: their papers show how the engines you use work.

Every date in this guide was read on the organiser's own page on 3 November 2026. Dates and venues change, and this page quotes no ticket prices, so confirm on the organiser's page before you book.

Start from the stack you run, because the value of a talk is how much of it you can use on Monday. If your warehouse is Snowflake, its summit shows the roadmap for the features you depend on and has talks from teams on the same platform. The same goes for Databricks at Data + AI Summit and for dbt at dbt Summit. Pick the one that matches your largest bill.

If your work is streaming or orchestration, follow the tool rather than the vendor. Current is the Kafka event and Airflow Summit is the Airflow event, and neither showed a 2027 date when checked. A team that runs Apache Iceberg tables should watch Iceberg Summit, which is the project community's own gathering.

If you want depth that outlives a product cycle, read the research conferences' proceedings even when you skip the trip. Papers on query optimisation and stream processing explain why the engines you use behave as they do, which is the kind of reasoning a data pipeline interview tests when it asks you to justify a design instead of name a tool.

If you want to meet people rather than watch sessions, the independent events (AI Council, Data Engineering Summit, Big Data LDN) tend to put the hallway first. Vendor events are the largest and the hardest to get a quiet conversation at.

Current, Confluent's data streaming event, listed London in May 2026 and San Francisco on 4 to 5 November 2026, with no 2027 edition on its site when checked. Airflow Summit's 2026 edition ran 31 August to 2 September 2026 in Austin, and its page mentions no 2027 edition. Iceberg Summit's 2026 edition was 2 in-person days on 8 and 9 April 2026 in San Francisco, announced on the Apache Iceberg site, with no 2027 edition read.

AWS re:Invent 2026 runs 30 November to 4 December 2026 in Las Vegas, and its 2027 dates were not on the page when read. The pattern across these events is yearly, so plan for a 2027 edition and wait for the organiser to confirm it instead of booking against a guessed date. A calendar hold costs nothing, and travel booked on a rumoured date can cost a lot.

Most organisers publish recordings, and that is often the better use of your time. The dbt Summit page keeps its 2026 keynotes and breakouts on demand, Airflow Summit keeps its 2026 sessions on demand and ran a free online Airflow Summit Online Reconnect on 4 and 5 November 2026, and Snowflake's summit page offers the keynotes and a set of popular sessions from the previous edition. Iceberg Summit 2026 offered a free streaming pass for the keynotes and 1 track.

What you give up is the hallway: the conversation after a talk where a maintainer tells you the limit that the slides left out. If your budget allows 1 trip, spend it on the event closest to the tool that causes your worst incidents, and watch the rest as recordings.

Decide before you arrive which 2 or 3 problems you want an outside view on, such as a backfill that takes too long, a schema change process that breaks consumers, or a cost line that grows faster than data volume. Then choose sessions by those problems, not by speaker name, and ask every speaker the same question: what failed, and what did you stop doing?

Customer talks at vendor events are strongest when the speaker gives scale and a failure. Treat a talk with neither as marketing and leave. Write one page for your team within a week: the 3 things you would change, the 1 you would not, and the tool you ruled out.

Conference talks also feed interview preparation. Design rounds ask about the topics the talks cover, such as exactly-once delivery, late data and table formats. Practise explaining a design with the trade-offs out loud on data pipeline practice problems, and see the Kafka interview questions if you came back from Current with gaps. The full path is in the data engineer interview prep guide.

A conference fills a few days of a year. The rest of the year, the same ideas arrive through communities and newsletters, which cost nothing. The data engineering community guide lists where practitioners gather, and the data engineering newsletters guide covers the reading worth keeping.