Key takeaways
- Pick by the layer of the stack you are deciding on: PyTorch Conference and MLSys for training and serving, KubeCon and Ray Summit for the platform underneath, Data + AI Summit for the data layer.
- Two clashes to plan around: Data + AI Summit (21–24 June) and MLSys (21–25 June) share a week, and March stacks KubeCon Europe (15–18 March) against AI Council in San Francisco (23–24 March).
- For practitioner talks with little vendor pitching, QCon London (13–15 April) and AI Council are the strongest picks; ODSC AI East (10–12 May) is the one built around hands-on training.
- Ray Summit, MLOps World, ODSC AI West and Snowflake Summit had not published 2027 dates when I checked, so budget for them as placeholders.
When are the AI engineering conferences in 2027?
| Dates | Event | Where |
|---|---|---|
| 9–11 Feb 2027 | AI DevWorld | Santa Clara · Santa Clara Convention Center |
| 15–18 Mar 2027 | KubeCon + CloudNativeCon Europe | Barcelona · Fira Gran Via |
| 23–24 Mar 2027 | AI Council | San Francisco · Palace Hotel |
| 13–15 Apr 2027 (certification day 16 Apr) | QCon London | London · QEII Centre |
| 26–27 Apr 2027 | AI Engineer Miami | Miami · Hyatt Regency Miami |
| 10–12 May 2027 | ODSC AI East | Boston · Boston Convention and Exhibition Center |
| 21–24 Jun 2027 | Data + AI Summit | San Francisco · Moscone North, West and South (and virtual) |
| 21–25 Jun 2027 | MLSys | Bellevue, Washington |
| TBA (2026 ran 24–26 Aug) | Ray Summit | San Francisco (2026 edition: Marriott Marquis) |
| 6–7 Oct 2027 | PyTorch Conference | San Jose, California |
Which AI engineering conferences are worth attending?
Ranked by how much a technology leader gets out of the trip. "Checked against" says how firm each date is: an organiser's own site beats event press and venue listings, and both beat third-party calendars.
-
6–7 Oct 2027 · San Jose, California
For: Leads who own model training, fine-tuning or inference platforms
The PyTorch Foundation's annual conference, run by the Linux Foundation, and the main yearly gathering for the people who maintain and build on the framework. Send whoever owns your training or serving platform, since this is where you hear what the maintainers plan before it ships. The venue and programme were not yet published.
Checked against: Organiser site (Linux Foundation events page)
-
15–18 Mar 2027 · Barcelona · Fira Gran Via
For: Platform and infrastructure teams running models on Kubernetes
The CNCF's European flagship, with 13,000+ attendees and 350+ sessions by the organiser's count, and an explicit focus on the infrastructure behind AI systems. The co-located Cloud Native AI + Inference Day on 15 March covers LLM inference, RAG, MLOps and moving AI from development into production. The North American edition follows on 8–11 November in New Orleans.
Checked against: Organiser site
-
21–24 Jun 2027 · San Francisco · Moscone North, West and South (and virtual)
For: Data platform owners, especially Databricks customers
Databricks' annual conference, spread across three Moscone halls, and the biggest single event for teams whose models sit on top of a lakehouse. Go if Databricks runs your data platform, because the platform roadmap you will build on next year is announced here. It runs the same week as MLSys, so split the team rather than the budget.
Checked against: Organiser site
-
23–24 Mar 2027 · San Francisco · Palace Hotel
For: Staff engineers, data and AI infrastructure leads, technical founders
A deliberately technical conference that draws 1,200+ engineers, founders and researchers, by the organiser's count, with past speakers such as DuckDB co-founder Hannes Mühleisen. Small enough to talk to the people who build the data and inference infrastructure you buy. Early speaker applications were open when I checked.
Checked against: Organiser site
-
13–15 Apr 2027 (certification day 16 Apr) · London · QEII Centre
For: Engineering managers, architects and senior engineers in Europe
InfoQ's practitioner conference, with 15 tracks and a rule of patterns over pitches. The 2027 programme includes tracks on eval platforms and guardrails for AI systems, architecture in the age of AI, modern data platforms and managing teams through AI adoption. Early-bird pricing was £2,490 until 8 December 2026.
Checked against: Organiser site
-
21–25 Jun 2027 · Bellevue, Washington
For: Research leads and ML infrastructure architects
The tenth Conference on Machine Learning and Systems, an academic venue for research at the boundary of models and hardware: training efficiency, inference serving, compilers and scheduling. Expect peer-reviewed papers, some of which become the serving and training techniques vendors sell a few years later. Paper notifications go out on 28 February 2027.
Checked against: Organiser site (it also says registration and schedule details are still being confirmed)
-
10–12 May 2027 · Boston · Boston Convention and Exhibition Center
For: Engineering teams that need hands-on training, plus their managers
The Open Data Science Conference has run since 2015 and bills itself as the hands-on conference for AI builders. Tracks include AI engineering, evals and AIOps, data and AI infrastructure and agentic systems, and the AI X Leadership Summit covers how organisations build, govern and scale AI. Useful when the goal is upskilling a team, less so for peer networking at executive level.
Checked against: Organiser site
-
9–11 Feb 2027 · Santa Clara · Santa Clara Convention Center
For: Developers adopting AI tools and the managers who budget for them
Co-located with DeveloperWeek, it reports 6,000 total participants and 100+ exhibitors. Tracks include AI and ML engineering, AI ops, infrastructure and MCP, AI security and an AI Executive Summit. The earliest event on this list, and a broad survey of developer AI tooling with a large expo.
Checked against: Organiser site
-
TBA (2026 ran 24–26 Aug) · San Francisco (2026 edition: Marriott Marquis)
For: Teams scaling training, batch inference or data processing on Ray
Anyscale's conference for the Ray distributed compute framework, which many teams use to scale training and inference across clusters. Worth a slot if Ray is in your stack or on your shortlist. The 2027 edition had not been announced when I checked.
Checked against: Organiser site (2027 dates not yet announced)
-
26–27 Apr 2027 · Miami · Hyatt Regency Miami
For: AI engineers, CTOs and VPs of AI in the US South-East and Latin America
A regional edition of the AI Engineer series, organised by G2i and the team behind React Miami as part of Frontier Tech Week. About 500 attendees and around 24 talks, and the organiser says a past edition was roughly 70% decision makers. A compact option if the World's Fair is too large or too far.
Checked against: Organiser site
How do I choose between the AI engineering conferences?
Start from the layer of the stack where you have a decision pending. If it is the model layer, such as which serving engine to standardise on or whether to fine-tune in-house, PyTorch Conference and MLSys are where the techniques are presented by the people who wrote them. If it is the platform, such as running inference on Kubernetes or scaling batch jobs, send the platform lead to KubeCon Europe with a day at Cloud Native AI + Inference Day, or to Ray Summit if Ray is already in use. For a primer on the serving options before anyone travels, read Ollama vs LM Studio vs llama.cpp vs vLLM.
If the decision sits in the data layer, the vendor conferences matter more than their reputation suggests. Data + AI Summit is where Databricks customers learn what they will be migrating to next, and Snowflake has announced Summit 27 without dates yet. For the practice of shipping LLM features, as opposed to the infrastructure, QCon London and AI Council give the most production stories per hour. Agent-specific events are on the separate agentic AI conferences list, which this page avoids repeating.
How should I plan the 2027 calendar and budget?
Plan around two clashes. Data + AI Summit and MLSys both start on 21 June, and March puts KubeCon Europe in Barcelona one week before AI Council in San Francisco. Most teams should pick one per cluster and send a different person to each, which also spreads the conference load away from the same two senior engineers every year.
Hold budget for the events that had not published 2027 dates on 19 November 2026: Ray Summit, MLOps World (its 2026 edition ran 17–18 November in Austin), ODSC AI West and Snowflake Summit. Ask each attendee for a written readout that names one change to your platform, cost model or hiring plan. If the readout cannot point to one, the event is a candidate to watch on video next year, and the money is better spent on the inference bill, which the LLM access cost calculator helps size.
Running a AI engineering conference?
Send corrections or a missing event to [email protected]. We add events editorially when they fill a gap in the list. Organisers who want a featured, labelled Sponsored slot at the top of this page can ask about a listing.
Related questions
More AI conference lists for 2027
- AI Conferences 2027: the full calendar
- Best Agentic AI Conferences in 2027
- Best Agentic Marketing Conferences in 2027
- Best AI Marketing Conferences in 2027
- Best GEO and AI Search Conferences in 2027
- Best AI Chip and Semiconductor Conferences in 2027
- Best AI Data Center Conferences in 2027
- Best AI Governance Conferences in 2027
- Best AI in Healthcare Conferences in 2027
- Best AI Security Conferences in 2027
- Best Physical AI and Robotics Conferences in 2027
- Best AI Customer Experience Conferences for Executives in 2027
- Best AI in Finance Conferences for Technology Leaders in 2027
- Best Enterprise AI Conferences for CIOs and CTOs in 2027
- CTO Conferences 2027
- AI Education: where technology leaders learn AI in 2027