Sepid Hosseini (@sepidshs) header
Sepid Hosseini (@sepidshs) profile picture

ML Researcher @ RBC Borealis | PhD CS @SFU |@SeoulNatlUni Alumni | Computer Vision & FinTech |representation learning, optimization | https://xitter.gq/t.co/3rr0fJ8Zxu

Greater Vancouver A
Joined December 2022
Photo by Sepid Hosseini (@sepidshs)Photo by Sepid Hosseini (@sepidshs)Photo by Sepid Hosseini (@sepidshs)Photo by Sepid Hosseini (@sepidshs)
Still processing this, CAO v1 at ICLR 2026 was a huge success 🚀 So grateful to all speakers, panelists, reviewers, authors & participants ❤️ CAO v2 is coming (end of 2026!) Apply as reviewer: lnkd.in/gx2Pt-VW Papers 👇 sites.google.com/view/iclr-2…
3
113
⏰ Last 7 Days! ICLR 2026 Workshop “Catch, Adapt, and Operate: Monitoring ML Models Under Drift” Deadline Approaching 📅 Deadline: February 10, 2026 We’re looking for work on: • Drift detection & monitoring • Temporal / online / continual learning • Reliability & robustness under non-stationarity • Human-in-the-loop systems • Real-world failures & lessons learned 🏆 Best Paper & Best Poster Awards ✈️ Limited Travel grants available Whether your work is theoretical, applied, or even study of methods that are not working, this is the kind of venue we want to create. 🌍 ICLR 2026, Rio de Janeiro sites.google.com/view/iclr-2… Don’t wait until the last minute, 7 days goes fast.
📣 Call for Papers, ICLR 2026 Workshop Catch, Adapt, and Operate (CAO): Monitoring ML Models Under Drift As ML systems move from benchmarks to real-world deployment, distributional drift becomes unavoidable. This workshop brings together researchers and practitioners working on how to monitor, adapt, and reliably operate machine learning models in dynamic environments. 📄 Submission Tracks To encourage a broad range of contributions, CAO offers multiple tracks: 🔹 Main Track Full technical papers with solid methodology and experimental validation. 🔹 Tiny Papers Track Shorter papers for early ideas, focused insights, or promising preliminary results. 🔹 Special Track, Lessons from Failures A dedicated space for negative results, failed experiments, and practical lessons that don’t always fit traditional venues , but are critical for progress in real-world ML. Paper submissions now open ! 📍 Rio de Janeiro, Brazil 🔗 sites.google.com/view/iclr-2…… #ICLR2026 #MachineLearning #MLSystems #ModelMonitoring @iclr_conf #ICLRworkshop #iclr2026
2
196
📣 Call for Papers, ICLR 2026 Workshop Catch, Adapt, and Operate (CAO): Monitoring ML Models Under Drift As ML systems move from benchmarks to real-world deployment, distributional drift becomes unavoidable. This workshop brings together researchers and practitioners working on how to monitor, adapt, and reliably operate machine learning models in dynamic environments. 📄 Submission Tracks To encourage a broad range of contributions, CAO offers multiple tracks: 🔹 Main Track Full technical papers with solid methodology and experimental validation. 🔹 Tiny Papers Track Shorter papers for early ideas, focused insights, or promising preliminary results. 🔹 Special Track, Lessons from Failures A dedicated space for negative results, failed experiments, and practical lessons that don’t always fit traditional venues , but are critical for progress in real-world ML. Paper submissions now open ! 📍 Rio de Janeiro, Brazil 🔗 sites.google.com/view/iclr-2…… #ICLR2026 #MachineLearning #MLSystems #ModelMonitoring @iclr_conf #ICLRworkshop #iclr2026
1
8
2
26
30,747
Sepid Hosseini retweeted
Proud to announce the first open source time-series foundation model Llama (inspired by @Meta's LLaMA - thanks, @ylecun ;) , built jointly with my awesome collaborators from @cerc_aai @UMontreal @ServiceNowRSRCH and @MorganStanley: @arjunashok37 @krasul, @CluelessAndrew, @hena_ghonia, @rishika2110, @Arian_Khorasani, @653mjd, @GeorgeAdams101, Roland Riachi, Nadhir Hassen, Marin Bilos, Sahil Garg, Anderson Schneider, @NicolasChapados, @alexandredrouin, @vzantedesc, Yuriy Nevmyvaka.
Foundation models are well-established in vision and language, but time series forecasting has lagged behind - it still relies on dataset-specific models. Meet Lag-Llama: the first open-source foundation model for time series forecasting!
8
32
1
146
33,823
Sepid Hosseini retweeted
👋 Come work with us! Applications for our Summer 2024 #Engineering Co-op are now open. As a Co-op #ML #SoftwareEngineer, you’ll be involved in a project end to end – data pre-processing to implementing algos and front-end development. Apply by Jan 25 ➡️ bit.ly/3RQnj0A
2
4
683
Sepid Hosseini retweeted
BREAKING: @BorealisAI is a sponsor of @NeurIPSConf! Our researchers will be presenting #AI products, research, and career opportunities at our booth from Dec. 11-14. Check out our latest blog post at bit.ly/3uM5TJp to learn more. 👋 See you at #NeurIPS2023!
3
7
617
Exciting news! 🎉 Our paper, "Puzzlefusion: Unleashing the Power of Diffusion Models for Spatial Puzzle Solving" has been accepted at NeurIPS 2023 as a spotlight. #neurips @amin_shabani @YasutakaFuruka1 @NeurIPSConf
Diffusion Models are general-purpose learning machinery. We have explored the use for reconstruction and alignment estimation, achieving SOTA performance for HD-mapping, floorplan reconstruction, room arrangement, and jigsaw puzzle. poly-diffuse.github.io/ sepidsh.github.io/puzzlefusi…
2
15
3,042
Sepid Hosseini retweeted
OPEN ROLE: Join our team as a DevOps Engineer experienced in designing and supporting machine learning infrastructure and automation tools. Grow your career and work with a passionate team committed to bringing #ML to Enterprise. Apply today: grnh.se/949a5f982us.
1
1
4
601
Sepid Hosseini retweeted
We’re gearing up for the International Conference on Machine Learning, happening July 23-29, with some reading! If you (like us) care about #AI being developed responsibly, here are a few #ICML2023 papers on topics of #responsibleAI to add to your reading list 🧵:
1
2
8
1,608
Sepid Hosseini retweeted
HouseDiffusion: Vector Floorplan Generation via a Diffusion Model with Discrete and Continuous Denoising Mohammad Amin Shabani, @sepidshs , @YasutakaFuruka1 github.com/aminshabani/house… Poster 127 #CVPR2023
1
4
1
30
4,184
State-of-the-art Floorplan Generation directly in vector space via continuous and discrete denoising. This afternoon poster at 127 (now 4:30pm - 6:30pm) #CVPR2023 done by @amin_shabani and @sepidshs
4
29
3,627
Exciting news! @amin_shabani and I will present our paper “HouseDiffusion: Vector Floorplan Generation via a Diffusion Model with Discrete and Continuous Denoising” today afternoon @cvpr2023 if you are around please come and check it out and let’s have a talk. #cvpr2023
1
1
9
1,422
I stand with graduate students from @SFU_CompSci / @sfu in securing a fair wage for their (TA) work. As an associate faculty at #SFU and a SFUFA member, today I am cancelling my graduate class, and honour the @TSSU picket line. Best of luck
3
10
52
4,880
Diffusion Models are general-purpose learning machinery. We have explored the use for reconstruction and alignment estimation, achieving SOTA performance for HD-mapping, floorplan reconstruction, room arrangement, and jigsaw puzzle. poly-diffuse.github.io/ sepidsh.github.io/puzzlefusi…
3
82
6
393
73,737
Diffusion models can learn different kinds of arrangements, put the pieces together, and reconstruct the whole object, from jigsaw puzzles to shredded documents, from rooms in a house to Voronoi partitions. #puzzlefusion #diffusion #DL #AI sepidsh.github.io/puzzlefusi…
1
6
20
3,256