<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0"><channel><description>Ml Research Engineer at the intersection of model training and efficient inference on NPUs</description><link>https://bsky.app/profile/ramchalamkr.bsky.social</link><title>@ramchalamkr.bsky.social - Ramchalam K R</title><item><link>https://bsky.app/profile/ramchalamkr.bsky.social/post/3lz6rrastec2q</link><description>Excited to announce that our Paper from #Qualcomm Canada has been accepted at #NeurIPS2025 OmniDraft: A Cross-vocabulary, Online Adaptive Drafter for On-device Speculative Decoding.&#xA;Looking forward to sharing our work at NeurIPS 2025. @neuripsconf.bsky.social&#xA;Preprint - arxiv.org/abs/2507.02659</description><pubDate>19 Sep 2025 11:40 +0000</pubDate><guid isPermaLink="false">at://did:plc:r4suowvgg6zfo3zoeb2xi6if/app.bsky.feed.post/3lz6rrastec2q</guid></item><item><link>https://bsky.app/profile/ramchalamkr.bsky.social/post/3ld7z47cfru2k</link><description>Thanks everyone for the wonderful and engaging discussions during our poster sessions and live demo of our work - Stepping Forward On The Last Mile. Thank you 😊.&#xA;https://arxiv.org/html/2411.04036v1&#xA;&#xA;@neuripsconf.bsky.social &#xA;@qualcomm.bsky.social</description><pubDate>14 Dec 2024 00:09 +0000</pubDate><guid isPermaLink="false">at://did:plc:r4suowvgg6zfo3zoeb2xi6if/app.bsky.feed.post/3ld7z47cfru2k</guid></item><item><link>https://bsky.app/profile/ramchalamkr.bsky.social/post/3lcsma4svks27</link><description>Qualcomm at NeurIPS 2024: Our groundbreaking innovations and cutting-edge advancements in AI.&#xA;https://www.qualcomm.com/news/onq/2024/12/qualcomm-at-neurips-2024-ai-research-demos-papers&#xA;&#xA;If you&#39;re attending NeurIPS 2024 in Vancouver, be sure to visit us at Qualcomm&#39;s booth #533 and our poster(#6102) on Wednesday - Stepping forward on the last mile.</description><pubDate>08 Dec 2024 16:14 +0000</pubDate><guid isPermaLink="false">at://did:plc:r4suowvgg6zfo3zoeb2xi6if/app.bsky.feed.post/3lcsma4svks27</guid></item><item><link>https://bsky.app/profile/ramchalamkr.bsky.social/post/3lcnkaasnwk2t</link><description>Our work at Qualcomm AI research, Qualcomm Canada was accepted at NeurIPS 2024. This work presents on-device training for any network (transformer, convolutional or rnn architecture) using fixed points(quantized) forward gradients (no back-prop).&#xA;Paper link: https://arxiv.org/html/2411.04036v1&#xA;&#xA;#Neurips2024</description><pubDate>06 Dec 2024 15:55 +0000</pubDate><guid isPermaLink="false">at://did:plc:r4suowvgg6zfo3zoeb2xi6if/app.bsky.feed.post/3lcnkaasnwk2t</guid></item></channel></rss>