/
Meeting notes

Meeting notes

Meeting Info

AI/ML Framework meetings are open to the public and held on Tuesdays at noon UTC.  World Time Zone Map

Zoom 1 Info:

UTC-time

Tuesday 12:00 (Summer)

Tuesday 13:00 (Winter)

https://zoom.us/j/9644759813

Seoul (KST / UTC+9)

Tuesday 21:00

Tuesday 22:00

TODO

Delhi (IST / UTC+5:30)

Tuesday 17:30

Tuesday 18:30

TODO

Seattle (PST / UTC -8)

Tuesday 04:00





New York (EDT / UTC -4 during DST (-5 in the winter))

Tuesday 08:00

Tuesday 09:00

+1 669 900 6833 // +1 646 558 8656 // 855 880 1246 (Toll Free) // 877 369 0926 (Toll Free)

Beijing (CST / UTC+8)

Tuesday 20:00

Tuesday 21:00

+86 10 87833177 // +86 10 53876330 // 400 616 8835 (Toll Free) // 400 669 9381 (Toll Free)

Tokyo (JST / UTC+9) 

Tuesday 21:00

Tuesday 21:00

+81 524 564 439 // +81 3 4578 1488 // 0 800 100 5040 (Toll Free)

Other international numbers

World Time Zone Map



International numbers available: https://zoom.us/u/acyy3hylQi

 

All meeting notes

2025-11-25 (Tuesday) : Next Meeting

2025-11-11 (Tuesday) :

✅ Completed Items

  • AIMLFW-300 – Upgrade Containerd, Nerdctl, and Buildkit versions compatible with Kubernetes v1.32.8 (Ashish Jain).

  • AIMLFW-299 – Upgrade Calico version compatible with Kubernetes v1.32.8 (Ashish Jain).

  • AIMLFW-298 – Upgrade Kubernetes version to v1.32.8 (Ashish Jain).

  • AIMLFW-289 – Update AIMLFW pipeline to use MM SDK version 0.4 (Swaraj Kumar).

  • AIMLFW-287 – Validate Kubeflow with TensorFlow 2.17.1 (Moksh Baweja).

🔄 In Progress

  • AIMLFW-297 – Add pipeline list table to Training Function page (YongHyun Kwon).

  • AIMLFW-290 – Fix model upload ambiguity by using model ID and adding type argument in MME upload (Swaraj Kumar).

  • AIMLFW-282 – Upgrade MME model-upload to support Artifact-Version update (Ashish Jain).

  • AIMLFW-276 – Align MME uploadModel function and Model Storage SDK keys for consistent upload/download (Swaraj Kumar).

  • Github actions enabled for MME, TM and data-extraction repository.

📝 To Do

  • AIMLFW-295 – Beautify AIMLFW installation logs (Ashish Jain).

  • AIMLFW-293 – Expose LSTMModel configuration parameters in model_factory (Subhash).

  • AIMLFW-292 – Remove hardcoding of featureList; derive dynamically from featurepath parameter (Subhash).

  • AIMLFW-283 – Deprecate UpdateArtifact endpoint in favor of standard UpdateModel (Subhash).

  • AIMLFW-281 – Remove redundant/confusing model existence check (Ashish Jain).

  • AIMLFW-280 – Add unit tests for pipeline_controller.py (Kangmin Han).

  • AIMLFW-277 – Parameterize temporary dataset file path in model training component (Sumin Kang).

2025-10-28 (Tuesday) : Cancelled

2025-10-14 (Tuesday) :

✅ Completed Items

  • AIMLFW-294 – Updated GitHub Actions tag patterns to require 3-segment version numbers (Swaraj).

  • AIMLFW-286 – Listed patches required for L-Release stability (Moksh Baweja).

  • AIMLFW-279 – Fixed InfluxDB installation issue (Moksh Baweja).

  • AIMLFW-278 – Fixed Cassandra installation issue (Moksh Baweja).

🔄 In Progress

  • AIMLFW-300 – Upgrade Containerd, Nerdctl, and Buildkit versions compatible with Kubernetes v1.32.8 (Ashish Jain).

  • AIMLFW-299 – Upgrade Calico version compatible with Kubernetes v1.32.8 (Ashish Jain).

  • AIMLFW-298 – Upgrade Kubernetes version to v1.32.8 (Ashish Jain).

  • AIMLFW-297 – Add pipeline list table to Training Function page (YongHyun Kwon).

  • AIMLFW-290 – Fix model upload ambiguity by using model ID and adding type argument in MME upload (Swaraj Kumar).

  • AIMLFW-289 – Update AIMLFW pipeline to use MM SDK version 0.4 (Swaraj Kumar).

  • AIMLFW-287 – Validate Kubeflow with TensorFlow 2.17.1 (Moksh Baweja).

  • AIMLFW-282 – Upgrade MME model-upload to support Artifact-Version update (Ashish Jain).

  • AIMLFW-276 – Align MME uploadModel function and Model Storage SDK keys for consistent upload/download (Swaraj Kumar).

📝 To Do

  • AIMLFW-295 – Beautify AIMLFW installation logs (Ashish Jain).

  • AIMLFW-293 – Expose LSTMModel configuration parameters in model_factory (Subhash).

  • AIMLFW-292 – Remove hardcoding of featureList; derive dynamically from featurepath parameter (Subhash).

  • AIMLFW-283 – Deprecate UpdateArtifact endpoint in favor of standard UpdateModel (Subhash).

  • AIMLFW-281 – Remove redundant/confusing model existence check (Ashish Jain).

  • AIMLFW-280 – Add unit tests for pipeline_controller.py (Kangmin Han).

  • AIMLFW-277 – Parameterize temporary dataset file path in model training component (Sumin Kang).

⚡ Highlights

  • Infrastructure upgrade to Kubernetes v1.32.8 and related components (Containerd, Calico) is underway — critical for upcoming M-release stability.

  • SDK and MME alignment progressing with tasks (AIMLFW-289, 290, 282, 276) to ensure seamless model upload and versioning.

  • Kubeflow validation with TensorFlow 2.17.1 (AIMLFW-287) nearing completion, ensuring compatibility for L/M releases.

  • UI enhancement in progress for Training Function page (AIMLFW-297).

  • Logging and configuration flexibility improvements planned (AIMLFW-292, 293, 295).

🎯 Action Items

  • Prioritize completion of K8s ecosystem upgrades (AIMLFW-298, 299, 300) to unblock downstream validation.

  • Finalize SDK alignment and model upload consistency before regression testing (AIMLFW-289, 290, 282, 276).

  • Complete Kubeflow validation with TF 2.17.1 (AIMLFW-287) for end-to-end workflow testing.

  • Proceed with log beautification and dynamic parameterization to enhance maintainability (AIMLFW-295, 292).

  • Review pending To-Do items for prioritization before the next sprint cycle.

2025-09-30 (Tuesday) :

✅ Completed Items

  • AIMLFW-271 – Documentation of Pipeline Component (Moksh Baweja).

  • AIMLFW-266 – Added export_model functionality in mm_sdk (Swaraj Kumar).

  • AIMLFW-259 – Upload Pipeline Script (Moksh Baweja).

  • AIMLFW-258 – Fixed missing kfp-kubernetes dependency (Moksh Baweja).

  • AIMLFW-248 – Updated PostgreSQL image in installation (Swaraj Kumar).

  • AIMLFW-246 – Fixed retrain failure with TF 2.20.0 (Swaraj Kumar).

  • AIMLFW-245 – Resolved Helm source unavailability (Moksh Baweja).

  • AIMLFW-244 – Fixed pipeline build failure due to FeatureStore SDK (Ashish Jain).

  • AIMLFW-243 – Resolved version conflict between Cassandra Driver and FeatureStoreSDK (Subhash).

  • AIMLFW-242 – Fixed build failure of AIML pipeline components (Moksh Baweja).

🔄 In Progress

  • AIMLFW-274 – Container tag version bump to 5.0.0 (Swaraj Kumar).

  • AIMLFW-272 – Training job restart failure after pipeline upload issue (Swaraj Kumar).

  • AIMLFW-268 – Make recipe name configurable in install_traininghost.sh (Moksh Baweja).

  • AIMLFW-267 – Spec-compliant Training Manager + MME.

  • AIMLFW-261 – MME uploads model even if not registered (Ashish Jain).

  • AIMLFW-260 – install_leofs.sh Helm chart values file hardcoding (Moksh Baweja).

  • AIMLFW-257 – Refactor pipelines to use MME for model uploads/downloads (Swaraj Kumar).

  • AIMLFW-252 – Remove kserve adapter from recipe examples (Swaraj Kumar).

  • AIMLFW-251 – Update Dockerfile to use versioned SDKs (Moksh Baweja).

  • AIMLFW-250 – Platform compatibility matrix & upgradation (Ashish Jain).

  • AIMLFW-247 – Align AIMLFW with Ubuntu 24.04 (Ashish Jain).

⚡ Highlights

  • Major pipeline and SDK stability issues (AIMLFW-244, 243, 242) have been resolved.

  • High-priority functional gaps remain: Training Manager/MME spec compliance (AIMLFW-267), pipeline refactor (AIMLFW-257), and model upload bug (AIMLFW-261).

  • Infrastructure upgrades in progress: K8s version (AIMLFW-275), Ubuntu 24.04 alignment (AIMLFW-247), container tag bump (AIMLFW-274).

🎯 Action Items

  1. Expedite fixes for training job restart (AIMLFW-272) and model upload validation (AIMLFW-261).

  2. Coordinate with release team for container tag updates (AIMLFW-274) before M release freeze.

  3. Review spec-compliance progress (AIMLFW-267) to avoid slippage.

  4. Track infra upgrade dependencies: K8s version update (AIMLFW-275) and Ubuntu 24.04 compatibility (AIMLFW-247).

2025-09-16 (Tuesday) :

2025-09-02 (Tuesday) : Cancelled

2025-08-05 (Tuesday) : Cancelled

2025-07-22 (Tuesday) :

2025-08-19 (Tuesday) :

 

2025-07-08 (Tuesday) :

  • Recording Link: Not Available

  • Topics:

    • L Release : "L" Release

    • M Release Planning:

      • Model Deployment as rApp

      • Model inference Service

      • UAV path prediction pipeline based on pipeline-components

    • AIMLFW pipeline for generative AI : @Geon(Corbin) Kim

  • Action Items:

    • Review the design document for pipeline for generative AI.

2025-06-24 (Tuesday) : cancelled

2025-06-10 (Tuesday) :

Action Item:

  • @subhash proposed to @Ivan Seskar for deployment of AIMLFW to integration lab.

2025-05-27 (Tuesday) :

Action Item:

  • @subhash proposed to @Ivan Seskar for deployment of AIMLFW to integration lab.

2025-05-13 (Tuesday) :

Action Item:

  • @subhash proposed to @Ivan Seskar for deployment of AIMLFW to integration lab.

2025-04-29 (Tuesday) :

Action Item:

  • @subhash proposed to @Ivan Seskar for deployment of AIMLFW to integration lab.

2025-04-15 (Tuesday) :