KickstartAI ML engineer Sanne Wielinga is researching how to detect hallucinations in small, on-premise medical LLMs.
Public source
Publisher name
Public post
Our ML engineer Sanne Wielinga has been researching how to detect hallucinations in small, on-premise medical LLMs. This October, she’ll present the work at AIES 2026 in…
Company
KickstartAI
- Industry
- Software Development
- Location
- Delft, NL
- Company size
- 11–50 employees
About KickstartAI
AI can help society adapt and thrive. This is why at KickstartAI, we're driving the application of AI in the Netherlands where it adds the most value. Together with leading companies, we work on practical AI use cases that deliver measurable impact. From predicting emergency room patient influx to minimizing supermarket food waste and optimizing logistics operations, we apply AI where it makes a difference. Our projects, in partnership with key industry partners, bridge the gap between theory and real-world use. Simultaneously, we are building a thriving community of 7.500+ engineers, data scientists and AI experts connecting the brightest minds in AI. Our goal? To stimulate AI talent by helping them exchange ideas, best practices, and knowledge. Through these efforts, we're accelerating AI adoption in the Netherlands, ensuring the Dutch AI ecosystem not only survives but thrives in the rapidly evolving technological landscape.
See moreLatest activity
Latest activity from KickstartAI
11 signals
Partnerships
KickstartAI is partnering with ING to launch a new KickstartAI x ING Challenge focused on AI-powered fraud coaches for the Netherlands.
Sales & Marketing
KickstartAI achieved a 18% increase in sales and marketing over the previous year, driven by the collective support of over 100,000 Dutch customers and the collaboration with NPO Radio 2, VI, and Shownieuws.
Presence & Recognition
KickstartAI announced that early-bird tickets for TechDays 2026 are ending soon with attendees joining engineers, data scientists, and technical leaders at Circa Amsterdam on 15 October for a full day of applied AI.
Discover more
Similar signals
Similar public activity from other companies.
Research & Knowledge
H1
H1 published a new study testing LLMs on real-world clinical data to identify and assess patients against clinical trial eligibility criteria.
Research & Knowledge
Alloi AI
Alloi AI published a report tracing a third of healthcare AI agent rollbacks to a single failure point.
Research & Knowledge
distil labs
distil labs published a blog post titled the three ways a small model actually enters a production system, and how to tell which one you are looking at.
Research & Knowledge
Syndio
Syndio ran an existing medical answers dataset through two LLM judges, one with a patient rubric and another with a clinician rubric, revealing 53% disagreement rates and their reasons for disagreeing.
Research & Knowledge
Vals AI