{"id":14998,"date":"2026-01-09T18:34:01","date_gmt":"2026-01-09T15:34:01","guid":{"rendered":"https:\/\/teaduspark.ee\/startup-day-aire-club-44-multimodal-ai-combining-language-and-vision-for-better-results\/"},"modified":"2026-01-09T18:34:01","modified_gmt":"2026-01-09T15:34:01","slug":"startup-day-aire-club-44-multimodal-ai-combining-language-and-vision-for-better-results","status":"publish","type":"post","link":"https:\/\/teaduspark.ee\/en\/startup-day-aire-club-44-multimodal-ai-combining-language-and-vision-for-better-results\/","title":{"rendered":"(sTARTUp Day) AIRE Club #44: Multimodal AI &#8211; Combining Language and Vision for Better Results"},"content":{"rendered":"<p data-start=\"0\" data-end=\"47\"><img fetchpriority=\"high\" decoding=\"async\" class=\"alignnone wp-image-14980 size-full\" src=\"https:\/\/teaduspark.ee\/wp-content\/uploads\/2026\/01\/TTP-korvalurr-scaled.jpg\" alt=\"\" width=\"2560\" height=\"1347\" srcset=\"https:\/\/teaduspark.ee\/wp-content\/uploads\/2026\/01\/TTP-korvalurr-scaled.jpg 2560w, https:\/\/teaduspark.ee\/wp-content\/uploads\/2026\/01\/TTP-korvalurr-300x158.jpg 300w, https:\/\/teaduspark.ee\/wp-content\/uploads\/2026\/01\/TTP-korvalurr-1024x539.jpg 1024w, https:\/\/teaduspark.ee\/wp-content\/uploads\/2026\/01\/TTP-korvalurr-768x404.jpg 768w, https:\/\/teaduspark.ee\/wp-content\/uploads\/2026\/01\/TTP-korvalurr-1536x808.jpg 1536w, https:\/\/teaduspark.ee\/wp-content\/uploads\/2026\/01\/TTP-korvalurr-2048x1078.jpg 2048w\" sizes=\"(max-width: 2560px) 100vw, 2560px\" \/><\/p>\n<p data-start=\"0\" data-end=\"47\"><strong data-start=\"0\" data-end=\"47\">sTARTUp Day, Friday, 30.01 at 11 side-events room<\/strong><\/p>\n<p data-start=\"49\" data-end=\"172\"><strong>AIRE Club #44: Multimodal AI &#8211; Combining Language and Vision for Better Results<\/strong><\/p>\n<p data-start=\"174\" data-end=\"600\">Multimodal AI models offer strong potential, but today they often fall short in delivering the real-time accuracy needed for complex tasks like object counting or critical defect detection. This presentation examines why a single model is rarely sufficient, how combining language and vision models improves results, and which technical solutions and limitations shape their integration. <\/p>\n<p data-start=\"174\" data-end=\"600\">All the sTARTUp Day ticketholders are welcome, but please let us know you are joining us:  <a class=\"decorated-link\" href=\"https:\/\/tartuteaduspark.typeform.com\/to\/kalzF8Xo\" target=\"_new\" rel=\"noopener\" data-start=\"602\" data-end=\"650\">https:\/\/tartuteaduspark.typeform.com\/to\/kalzF8Xo<\/a><\/p>\n<p data-start=\"174\" data-end=\"600\"><strong>Speakers:<\/strong><\/p>\n<p data-start=\"652\" data-end=\"1061\"><strong data-start=\"652\" data-end=\"669\">Martin Rebane<\/strong><br data-start=\"669\" data-end=\"672\">Head of AI at Sparkup Tartu Science Park<br data-start=\"718\" data-end=\"721\"><br data-start=\"748\" data-end=\"751\">Martin has been developing and implementing practical AI systems since 2011. He earned a PhD in artificial intelligence from the University of Warwick in the UK. He possesses extensive hands-on experience in designing and deploying AI-based systems and processes for various organizations.  <\/p>\n<p data-start=\"1063\" data-end=\"1531\" data-is-last-node=\"\" data-is-only-node=\"\"><strong data-start=\"1063\" data-end=\"1082\">Ida Maria Orula<\/strong><br data-start=\"1082\" data-end=\"1085\">AI Developer at Sparkup Tartu Science Park<br data-start=\"1122\" data-end=\"1125\"><br data-start=\"1153\" data-end=\"1156\">Ida Maria holds a Master&#8217;s in Computer Science from the University of Tartu. In her studies, she specialized in health informatics and pharmacogenetics. 2021. In 2021, she joined the sTARTUp Day team\u2014first as a volunteer, later as the Program Manager. She is now working as an AI developer at the Sparkup Tartu Science Park. Ida Maria is a language enthusiast with a goal to speak 12 languages by 2030.   <\/p>\n","protected":false},"excerpt":{"rendered":"<p>sTARTUp Day, Friday, 30.01 at 11 side-events room AIRE Club #44: Multimodal AI &#8211; Combining Language and Vision for Better Results Multimodal AI models offer strong potential, but today they often fall short in delivering the real-time accuracy needed for complex tasks like object counting or critical defect detection. This presentation examines why a single [&hellip;]<\/p>\n","protected":false},"author":6,"featured_media":15036,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[194],"tags":[],"class_list":["post-14998","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-events"],"acf":[],"_links":{"self":[{"href":"https:\/\/teaduspark.ee\/en\/wp-json\/wp\/v2\/posts\/14998","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/teaduspark.ee\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/teaduspark.ee\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/teaduspark.ee\/en\/wp-json\/wp\/v2\/users\/6"}],"replies":[{"embeddable":true,"href":"https:\/\/teaduspark.ee\/en\/wp-json\/wp\/v2\/comments?post=14998"}],"version-history":[{"count":0,"href":"https:\/\/teaduspark.ee\/en\/wp-json\/wp\/v2\/posts\/14998\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/teaduspark.ee\/en\/wp-json\/wp\/v2\/media\/15036"}],"wp:attachment":[{"href":"https:\/\/teaduspark.ee\/en\/wp-json\/wp\/v2\/media?parent=14998"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/teaduspark.ee\/en\/wp-json\/wp\/v2\/categories?post=14998"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/teaduspark.ee\/en\/wp-json\/wp\/v2\/tags?post=14998"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}