#74 - PhD, ML/LLM, RAG, Opensource & CTO with Markus Stoll
🎙️ Who is Markus Stoll?
Markus Stoll is an expert AI guest who is well-versed in creating advanced software solutions for the efficient exploration of unstructured data. Markus is the Co-Founder and CTO of Renumics, a company that specializes in developing open-source tools for visualization in Data-Centric AI. Not only has Markus extensive experience in solving complex problems using workflow and visualization, but he also has a strong engineering background too. From Engineer to CTO, Markus has taken his passion for AI and founded Renumics, which has evolved over the years, and today, it's successfully developing machine learning and visualization solutions.
💡 In this episode...
... we discuss Markus's experience in creating advanced software solutions for exploring unstructured data. Markus discusses how feedback, domain understanding, and foundational models can be used to build scalable data evaluation projects and recommendation systems. He also highlights the challenges and opportunities of building RAG systems, the importance of glossaries and future fine-tuning, and introduces Spotlight, a powerful open-source tool for comprehensive data analysis. Finally, Markus introduces an open-source audio data segmentation and labeling tool that shifts from 3D visualization to audio data labeling based on user feedback.
⏳ Chapters
00:00:00 - Intro
00:00:47 - Efficient Problem Solving with Workflow and Visualization
00:02:22 - From Engineer to CTO and Co-founder of Renumics
00:04:54 - Markus' Journey in the AI Industry
00:14:51 -Renumics's CTO on Machine Learning and Engineering Challenges
00:15:40 - Building Scalable Data Evaluation Projects and Recommendation Systems
00:20:35 - Building and Solving Problems with Recommendation Systems
00:24:22 - Improving Recommendation Systems with User Feedback and Domain Understanding
00:25:50 - Improving RAG Systems with Data Visualization and Foundational Models
00:34:04 - Using Closed and Open Source Models for RAG Systems
00:36:29 - Scaling RAG Systems for Large Language Models
00:39:14 - Overcoming Challenges in Applying AI to Industry Data
00:47:46 - Spotlight: An Open Source Tool for Versatile Data Visualization
00:49:54 - Open Source Audio Data Segmentation and Labeling Tool
00:52:27 - Shift from 3D Visualization to Audio Data Labeling based on User Feedback
Markus Stoll: / markus-stoll-b39a42138
Markus's Medium: / markus.stoll
Github spotlight: https://github.com/Renumics/spotlight
Renumics: https://renumics.com/
Thomas Bustos: / thomasbustos
Artistic Direction & Video: Maxence Kerhoas
Follow Let's Talk AI:
✉️ Newsletter 👉 http://eepurl.com/ijZ8qz
🎙️ Podcast 👉 http://smartlink.ausha.co/let-s-talk-ai/
📹 Youtube 👉 / @lets-talk-ai
📷 Instagram 👉 / lets_talk_ai
🎞️ TikTok 👉 / letstalkai
🌐 Website 👉 https://lets-talk-ai.com/