idalab

We are seeking three talented summer interns to join our team this year, focusing on ML/AI for Biotechnology and Digital Health Our three internship projects: 🏗️ Data Engineering: Scale up evidence extraction for massive amounts of scientific data 📚 AI Methodology: Develop cascading evidence extraction for drug target identification using large language models (LLMs) 🩺 AI Strategy: Simulate real-world dialysis patient scenarios to improve therapy and treatment outcomes Why intern with idalab? ▪ Experience: Real-world projects with experienced mentors ▪ Knowledge: Insights into AI and data transformation in life sciences and healthcare ▪Work Culture: Open-minded and collaborative work environment ▪A fair compensation Who are we looking for? ▪ You are currently enrolled in a Master's degree or a PhD program OR you have recently graduated ▪ Background in life science or healthcare ▪ 6-8 week availability starting September or October ▪ Bonus points: Python, Linux, and cloud tech (e.g., Google Cloud Platform) skills Apply now! Apply via our website until August 11th #AI #LifeSciences #Healthcare #Internship #OpenPositions

Large Language Models 🤖 need careful supervision and training to stay on track in drug discovery 🧬. Here are five ways to prevent hallucinations and get the best results. https://lnkd.in/d7t6NgMp #ai #drugdiscovery

Clinical trial success rates in Phase 2 are below 30% 💊 📉. So what if we could predict efficacy before going into the clinic? Marco Schmidt (biotx.ai) explains how causal AI 🤖 🧬 identifies the most promising drug candidates, and will ultimately change the blockbuster-driven pharma business model. 🗓️ Tuesday, July 16th 📍idalab office, Potsdamer Str. 68, 10785 Berlin 18:30 Doors open 19:00 Talk starts 19:45 Drinks 👋🏾 Everyone's welcome 👉🏾 Register at idalab.de/seminars 📫 Subscribe to our newsletter for priority invitation Abstract The challenge in drug development is that clinical efficacy cannot be predicted. Success rates, especially in Phase II, when efficacy is tested for the first time, are below 30 percent. Causal AI enables the prediction of clinical efficacy, thus making clinical development more cost-effective and changing the pharma business model. The pursuit of a single blockbuster molecule with annual sales of one billion dollars is being replaced by several molecules with lower revenue potential but equal profitability. About the speaker Marco studied biochemistry in Berlin and Tübingen and began exploring human genetics and machine learning at the University of Cambridge, UK, which led to the founding of biotx.ai GmbH. The company has developed a causal AI to simulate clinical trials, and this technology is now being made available to biotech and pharmaceutical companies. He is also known for his textbook 'Chemical Biology & Drug Discovery'. About the idalab Seminar The idalab Seminar is a bi-monthly event focused on AI in Life Science and Healthcare, hosted at our office. Invited talk followed by drinks, and ample opportunity to connect with the Berlin ecosystem. #causalAI #biotech #pharma #machinelearning #data #artificialintelligence #clinicaltrials Dr. Sven Jungmann Reuben Honigwachs Ralf Banisch Philipp Hacker Nucleate Germany Vision Health Pioneers Incubator European New School of Digital Studies Life Sciences Tech Network - Berlin

🔨 It’s hammer time – or is it? When you only have a hammer, everything looks like a nail. This can be limiting. Sometimes you need to put away your favourite tools and look without prejudice at the problem. In this hospital-based case study, we explain the factors at play in designing an automated system to reduce false-positive ICU alarms – while prioritising patient safety. https://lnkd.in/dfNXJfxV #ai #healthcare #design

Every breath we exhale 💨 contains vital data, composed of over 1000+ unique molecules. Making sense of this molecular fingerprint is 🤖 #AI’s next frontier, to unlock early detection of a wide range of diseases. Join us for the next idalab Seminar with Dr. Sven Jungmann (Halitus GmbH) to learn how his startup combines breakthrough laser technology with AI to teach computers how to smell. 🗓️ Tuesday, May 7th 📍idalab office, Potsdamer Str. 68, 10785 Berlin 18:30 Doors open 19:00 Talk starts 19:45 Drinks 👋🏾 Everyone's welcome 👉🏾 Few spots left 📫 Subscribe to our newsletter for priority invitation Register at idalab.de/seminars Abstract With every breath we take, we exhale over 1000 unique molecules. That means there’s data in the very air that we breathe. Having taught machines to ‘hear’ and to ‘see’, a sense of smell is AI’s next frontier, and it’s a vital one. The molecular fingerprint of a single breath can reveal early signs of diseases such as breast cancer and Parkinson’s, quickly, non-invasively and cost effectively. Yet accurately detecting breath molecules comes with a lot of technological challenges; and the race is on to find out which team gets there first. In his talk, Sven explains his breakthrough laser spectrometer technology and why hardware is hard; outlines his tactics for fundraising, and describes how a pivot towards explainable machine learning ultimately helped him to succeed. About the speaker Dr. Sven Jungmann, a medical doctor turned entrepreneur, is the founder and CEO of Halitus. Previously, he held various leadership positions in healthcare, including partner and Chief Medical Officer at FoundersLane and Chief Medical Officer at the first corporate venture of Helios. Sven is an active angel investor and recently published his second book, titled “Wie gesund wollen wir sein” (Mosaik). About the idalab Seminar The idalab Seminar is a bi-monthly event focused on Data & ML/AI in Life Science & Health, hosted at our office. Invited talk followed by drinks, and ample opportunity to connect with the Berlin ecosystem. Gregor Obernosterer Dr. Christoph Zindel #medtech #diagnostic #healthcare #machinelearning #data #artificialintelligence #deeptec

😦 Nur einer von 5000 präklinischen Kandidaten schafft es, 1-3 Milliarden US-Dollar Kosten, 10-17 Jahre Entwicklungszeit? Das muss schneller gehen! 😦   Diese Woche erfuhren wir in der ersten Folge einer neuen Reihe unseres kostenlosen, virtuellen BioM #Lunchtalk zum Thema „AI/ML for Biotechnology“ wie die Unternehmen Cortex Discovery GmbH und idalab GmbH mit #DeepLearning-Technologien und generativen KI- #Sprachmodellen die #Arzneimittelforschung beschleunigen wollen. Diesen Lunchtalk verpasst? Hier können Sie sich die Vorträge ansehen: ➡️ https://lnkd.in/etpdM22G   Dr. Ivan De Weber, PhD, CEO bei Cortex Discovery GmbH nutzt 2 Technologien, die die Erfolgsquoten präklinischer Kandidaten erhöhen soll:   1️⃣ Next-generation “QSAR” = Quantitative Structure Activity Relationship lernt aus Daten von Molekülen und ihren Eigenschaften und überträgt trainierte Vorhersagen von Eigenschaften auf das neue Molekül Dazu nutzt das Unternehmen sein 🕸️ “Graph Neural Net” - eine eigens kreierte Datenbank mit 260 Millionen Datenpunkten, um jegliche Art von Eigenschaft wie z.B. Löslichkeit, Permeabilität und Aktivität aus der Molekülstruktur vorherzusagen – und dies alles zur gleichen Zeit!   2️⃣ “Docking” kombiniert Quantenchemie und #DeepLearning, um die Simulation von Ligand-Protein-Interaktionen zu revolutionieren. Die Modellierung der Wechselwirkung zwischen Arzneimittel-Targets soll mit der eigens entwickelten Technologie 1000x schneller sein als bisherige Methoden.   ---   Einen anderen Ansatz nutzt das Beratungs- und Dienstleistungsunternehmen idalab GmbH, das #MedTech-, #Pharma- und #Biotech-Unternehmen beim Einsatz von KI berät.   Julian Beimes, Senior Associate, führte anschaulich zum Thema “#LargeLanguageModels for supporting drug discovery” hin und stellte KI auf die gleiche Stufe disruptiver medizinscher Technologien wie dem modernen Licht-Mikroskop und der Röntgen-Kristallographie, die (neue) wissenschaftliche Herausforderungen angehen kann.   Beimes verdeutlichte die Fähigkeit #GenerativeAI in einem „Ozean an Informationen“ nicht nur Informationen zu erschließen, sondern die Fähigkeit aus biomedizinischen Texten Schlussfolgerungen ableiten zu können.   Zur Unterstützung der #Arzneimittelforschung nutzt Idalab "Large Language Models" um eine „Tailored Text Mining Machine“ zu bauen, diese zu erden und Informationen auf dem Niveau eines geschulten Menschen zu extrahieren - in einer Größenordnung, um anschließend aus der relevanten Literatur eine „Map of evidence“ zu erstellen, die den relevanten Teil des Proteoms mit potenziellen #Targets zeigt.   🙏 Herzlichen Dank an die Referenten für diesen faszinierenden Blick in eine KI-gestützte #DrugDiscovery. 🙏  

Schneller Forschung - schneller Kandidaten - schneller neue Arzneimittel und Therapien! Wir laden Sie herzlich ein, zur 1️⃣ ersten Folge einer neuen Reihe unseres kostenlosen, virtuellen #BioM #Lunchtalk, zum faszinierenden Thema „#AI/ML for #Biotechnology“ am kommenden 📅 Mittwoch, 22. November von 🕛 12 bis ca. 13 Uhr. Hier registrieren: ➡️ https://lnkd.in/e-QfrWV4 Es stellen sich die Unternehmen 🏢 Cortex Discovery GmbH und 🏢 idalab GmbH vor. Nutzen Sie die Gelegenheit, von Experten auf diesem Gebiet zu lernen, Einblicke in deren innovative Ansätze zu bekommen und mit ihnen in Kontakt zu treten. Es sprechen: Dr. Ivan De Weber, PhD, CEO bei #CortexDiscovery GmbH 👨💻 Vortragstitel: „Cortex's Modern Machine Learning Achieves Unprecedented Predictive Accuracy on Par with Experimental Precision.“ 👉 Cortex Discovery stellt zwei bahnbrechende #DeepLearning- #Technologien vor, die die #Arzneimittelforschung auf die nächste Stufe heben. Die eine nutzt biologische Experimente und liefert Vorhersagen, die der experimentellen Präzision in therapeutischen Bereichen entsprechen. Die zweite kombiniert #Quantenchemie und Deep Learning, um die Simulation von Ligand-Protein-Interaktionen zu revolutionieren. Damit will Cortex die Erfolgsquoten präklinischer Kandidaten erhöhen und Arzneimittelforschungsprojekte schon heute revolutionieren.   Julian B.eimes, Senior Associate bei #idalab GmbH 👨💻 Vortragstitel: „Large Language Models for supporting drug discovery” 👉 idalab ist ein Beratungs- und Dienstleistungsunternehmen, das Unternehmen aus dem Bereich #Healthcare/ #LifeScience beim Einsatz von KI unterstützt - von der Strategie über die Implementierung bis hin zum aktiven Betrieb. Zu unseren Kunden zählen Unternehmen wie Roche, HotSpot Therapeutics, Inc. und BIOTRONIK. Im Anschluss an die Vorträge folgt jeweils eine kurze #Diskussionsrunde unter Einbeziehung der Zuhörenden durch Fragenstellungen über den F&A-Chat. #Arzneimittel #Kandidat #Forschung #Entwicklung #Wirkstoff #Präklinik #DrugDiscovery

How can we use generative AI 🤖 in life science 🧬 and healthcare 🏥 while preserving patient privacy, data security and compliance with regulation? Join us for the next idalab Seminar with Reuben Honigwachs (Google) to find out. 🗓️ Wednesday, November 22nd 📍idalab office, Potsdamer Str. 68, 10785 Berlin 18:30 Doors open 19:00 Talk starts 19:45 Drinks 👋🏾 Everyone's welcome 👉🏾 Few spots left 📫 Subscribe to our newsletter for priority invitation Register at idalab.de/seminars Abstract While the potential of Generative AI for applications in healthcare, pharma and biotech is immense, there are concerns related to privacy, security and regulatory compliance. Reuben Honigwachs, Cloud Architect at Google, will outline the most pressing challenges for the sector and show how they can be met within and beyond the Google Cloud Platform. About the speaker Reuben Honigwachs is a Cloud Architect working as a Customer Engineer at Google Cloud. He has a background in DevOps and supports Google’s large customers in the Life Science and Healthcare sector. About the idalab Seminar The idalab Seminar is a bi-monthly event focused on Data & ML/AI in Life Science & Health, hosted at our office. Invited talk followed by drinks, and ample opportunity to connect with the Berlin ecosystem. Torsten Laupp Stefan Grosch James Bannister #genai #cloud #lifescience #healthcare #machinelearning #data #artificialintelligence

With great power comes great instability? 🔥 Channeling the enthusiasm of large language models 🤖 into a useful tool for drug discovery 💊 requires more than just prompting. It takes a village of tools, data, and ... real people. Best practices 💊 Ground the model to avoid hallucination 💊 In doing so, always tailor to the specific biological context  💊 Enforce a structured output format to enable structured post-processing 💊 Set up a rigorous and efficient workflow to keep the precious human in the loop 💊 Take care of data privacy #drugdiscovery #llm #ai #machinelearning #biotech #genai #noblesseoblige

What if the key to a blockbuster drug 💊 was hidden in a footnote of a paper from 1997? Well, you probably wouldn't find it 💡. And that's where large language models 🤖 come to the rescue. With human-like understanding, LLM-based text mining tools can unearth crucial bits of information for drug discovery: 💊 Prioritising disease-relevant targets from screens 💊 Assessing target-related adverse event profiles 💊 Identifying biomarkers for a new target #ai #llms #drugdiscovery #drugdevelopment #biotech #chatgpt #chatgpt4 #screening #adverseevents #biomarkers #immunooncology #immunology