--- schema: 1 type: video title: 'Personalization in the Era of LLMs - Shivam Verma, Spotify' description: >- Spotify represents Ariana Grande and Bruno Mars as sequences of six tokens. The first two are shared because both are pop artists. The remaining tokens diverge to capture what makes each distinct. That is a Semantic ID, and it is how Spotify teaches open-weight LLMs to reason over a catalog of 100 million tracks the same way they reason over words. Shivam Verma from Spotify's AI foundation team walks through the three components they assembled to personalize LLMs at scale without full fine-tuning. User embeddings trained on streaming history across 750 million users form the base. Semantic IDs compress catalog vectors into tokens the model can autoregressively generate, predicting the next song or episode as the next token in a sequence. A soft tokenization layer projects a user's embedding directly into the LLM's token space, giving the frozen model a user-specific token to attend over. Podcast next-episode recommendations are already running on this stack in production. Speaker info: - https://x.com/kaffeinated - https://www.linkedin.com/in/shivam13verma --- Watch along with the corrected transcript, summary, timestamps and resources on this talk's official AIE page: https://ai.engineer/talks/5YSJEP0HWzM-personalization-in-the-era-of-llms Join our upcoming AI Engineer conferences (see https://ai.engineer landing page): - AIEi Shanghai, November 5–6 — the first for meeting top Chinese labs and engineers - AIE CODE in SF, November 10–12 — the top AI Coding conference returns - AIEi Sydney, December 7–8 — the first AIEi run alongside NeurIPS 2026 in Sydney --- canonical: >- https://liquidgold.cc/channel/aidotengineer/explore/video/personalization-in-the-era-of-llms-shivam-verma-spo-3270f7b929ae markdown: >- https://liquidgold.cc/channel/aidotengineer/video/personalization-in-the-era-of-llms-shivam-verma-spo-3270f7b929ae.md llms: null parent: 'https://liquidgold.cc/channel/aidotengineer.md' channel: aidotengineer updated: '2026-10-24T14:10:49.791Z' language: null children: [] videoId: 5YSJEP0HWzM durationSeconds: 1212 publishedAt: '2026-05-19T13:00:06.000Z' searchChunks: - ordinal: 0 text: null startSeconds: 26 endSeconds: 148 watch: 'https://liquidgold.cc/watch/5YSJEP0HWzM?t=26' heldReason: rights - ordinal: 1 text: null startSeconds: 159 endSeconds: 247 watch: 'https://liquidgold.cc/watch/5YSJEP0HWzM?t=159' heldReason: rights - ordinal: 2 text: null startSeconds: 258 endSeconds: 375 watch: 'https://liquidgold.cc/watch/5YSJEP0HWzM?t=258' heldReason: rights - ordinal: 3 text: null startSeconds: 258 endSeconds: 375 watch: 'https://liquidgold.cc/watch/5YSJEP0HWzM?t=258' heldReason: rights - ordinal: 4 text: null startSeconds: 258 endSeconds: 375 watch: 'https://liquidgold.cc/watch/5YSJEP0HWzM?t=258' heldReason: rights - ordinal: 5 text: null startSeconds: 386 endSeconds: 506 watch: 'https://liquidgold.cc/watch/5YSJEP0HWzM?t=386' heldReason: rights - ordinal: 6 text: null startSeconds: 386 endSeconds: 506 watch: 'https://liquidgold.cc/watch/5YSJEP0HWzM?t=386' heldReason: rights - ordinal: 7 text: null startSeconds: 386 endSeconds: 506 watch: 'https://liquidgold.cc/watch/5YSJEP0HWzM?t=386' heldReason: rights - ordinal: 8 text: null startSeconds: 516 endSeconds: 636 watch: 'https://liquidgold.cc/watch/5YSJEP0HWzM?t=516' heldReason: rights - ordinal: 9 text: null startSeconds: 516 endSeconds: 636 watch: 'https://liquidgold.cc/watch/5YSJEP0HWzM?t=516' heldReason: rights - ordinal: 10 text: null startSeconds: 745 endSeconds: 832 watch: 'https://liquidgold.cc/watch/5YSJEP0HWzM?t=745' heldReason: rights - ordinal: 11 text: null startSeconds: 842 endSeconds: 951 watch: 'https://liquidgold.cc/watch/5YSJEP0HWzM?t=842' heldReason: rights - ordinal: 12 text: null startSeconds: 734 endSeconds: 745 watch: 'https://liquidgold.cc/watch/5YSJEP0HWzM?t=734' heldReason: rights - ordinal: 13 text: null startSeconds: 745 endSeconds: 832 watch: 'https://liquidgold.cc/watch/5YSJEP0HWzM?t=745' heldReason: rights - ordinal: 14 text: null startSeconds: 745 endSeconds: 832 watch: 'https://liquidgold.cc/watch/5YSJEP0HWzM?t=745' heldReason: rights - ordinal: 15 text: null startSeconds: 842 endSeconds: 951 watch: 'https://liquidgold.cc/watch/5YSJEP0HWzM?t=842' heldReason: rights - ordinal: 16 text: null startSeconds: 1052 endSeconds: 1212 watch: 'https://liquidgold.cc/watch/5YSJEP0HWzM?t=1052' heldReason: rights - ordinal: 17 text: null startSeconds: 1052 endSeconds: 1212 watch: 'https://liquidgold.cc/watch/5YSJEP0HWzM?t=1052' heldReason: rights --- # Personalization in the Era of LLMs - Shivam Verma, Spotify Spotify represents Ariana Grande and Bruno Mars as sequences of six tokens. The first two are shared because both are pop artists. The remaining tokens diverge to capture what makes each distinct. That is a Semantic ID, and it is how Spotify teaches open-weight LLMs to reason over a catalog of 100 million tracks the same way they reason over words. Shivam Verma from Spotify's AI foundation team walks through the three components they assembled to personalize LLMs at scale without full fine-tuning. User embeddings trained on streaming history across 750 million users form the base. Semantic IDs compress catalog vectors into tokens the model can autoregressively generate, predicting the next song or episode as the next token in a sequence. A soft tokenization layer projects a user's embedding directly into the LLM's token space, giving the frozen model a user-specific token to attend over. Podcast next-episode recommendations are already running on this stack in production. Speaker info: - https://x.com/kaffeinated - https://www.linkedin.com/in/shivam13verma --- Watch along with the corrected transcript, summary, timestamps and resources on this talk's official AIE page: https://ai.engineer/talks/5YSJEP0HWzM-personalization-in-the-era-of-llms Join our upcoming AI Engineer conferences (see https://ai.engineer landing page): - AIEi Shanghai, November 5–6 — the first for meeting top Chinese labs and engineers - AIE CODE in SF, November 10–12 — the top AI Coding conference returns - AIEi Sydney, December 7–8 — the first AIEi run alongside NeurIPS 2026 in Sydney --- Canonical: https://liquidgold.cc/channel/aidotengineer/explore/video/personalization-in-the-era-of-llms-shivam-verma-spo-3270f7b929ae · Channel: [AI Engineer](https://liquidgold.cc/channel/aidotengineer.md) > Search chunk text is held until public-source provenance is verified. ## Document chunks > Missing: no stored document chunks are available.