Hirundo
Category: AI Safety
Hirundo is a Machine Unlearning platform that enables enterprises to find and surgically remove unwanted data (e.g., inaccuracies, bias, toxic content, private information) from trained AI models, without the need for complete retraining. Hirundo was founded in 2023. The company is led by Ben Luria. Based in Tel Aviv, Israel & London, UK. Team size: 11-50. Total funding raised: $8 million USD. Latest round: Seed. Key investors include Maverick Ventures Israel, SuperSeed, Alpha Intelligence Capital, Tachles VC, AI.FUND, Plug and Play Tech Center.
- Founded
- 2023
- Headquarters
- Tel Aviv, Israel & London, UK
- Team size
- 11-50
- Total funding
- $8 million USD
Value proposition
Dramatically reduces the time and cost of maintaining AI models by making them 'forget' harmful or unwanted data. This increases model accuracy, safety, and compliance while avoiding the massive expense of retraining from scratch.
Products and solutions
Hirundo Machine Unlearning Platform
Unique value
The first company dedicated to the commercial application of machine unlearning. Instead of filtering outputs, Hirundo surgically modifies the core AI model to 'forget' specific data, addressing the root cause of issues like bias and hallucinations.
Target customer
Enterprises in regulated industries (like finance, healthcare, defense) that deploy high-stakes AI models and require a way to fix, debug, and ensure the compliance of those models in a scalable and cost-effective manner.
Industries served
Financial Services, Healthcare, Defense, Technology
Technology advantage
Proprietary 'Machine Unlearning' algorithms that can pinpoint and remove hallucinations, bias, and toxic data directly from trained AI models. This avoids the need for costly, time-consuming retraining and preserves overall model performance.
How they differentiate
Hirundo's core differentiator is 'Machine Unlearning,' which surgically removes unwanted data from already-trained models. This is reportedly 45x faster and significantly cheaper than retraining models from scratch, which is the standard approach used by competitors.
Main competitors
Eightfold.ai, Pymetrics, HireVue
Key partnerships
Piloting its technology with multinational corporations in finance, healthcare, and defense., Joined HPE Unleash AI partner program (March 2026)., Partnership with Asgent to offer machine unlearning in Japan (April 2026)., Collaboration with Google DeepMind on Gemmaverse (2026).
Notable customers
Piloting with multinational corporations in finance, healthcare, and defense.
Major milestones
Successfully closed an $8 million seed funding round in June 2025., Pioneered 'Machine Unlearning' as a commercially viable solution for enterprise AI., Featured as a leading innovator in multiple tech publications for its novel approach to AI safety., Won first prize at RAISE Summit Startup Competition, beating 1,100+ AI startups (July 2025)., Recognized in OWASP Gen AI Security Solutions Landscape Q2 2026., Joined HPE Unleash AI partner program (March 2026)., Partnered with Asgent to bring machine unlearning to Japan (April 2026)., Featured by Google DeepMind on Gemmaverse (2026).
Growth metrics
Currently engaged in pilot programs with multinational corporations in the finance, healthcare, and defense sectors.
Market positioning
Pioneering the 'Machine Unlearning' category, positioning itself as a critical tool for enterprises that need to fix and maintain live AI models without incurring massive retraining costs. It targets high-stakes, regulated industries where AI accuracy, safety, and compliance are non-negotiable.
Geographic focus
Global, with a focus on regulated industries in North America and the EU.
Patents and IP
Information not publicly available.
About Ben Luria
Ben Luria is the CEO and Co-Founder of Hirundo. He is a Rhodes Scholar with a background in strategic planning, foreign relations, and public relations from his time in the IDF. He has a strong academic background with degrees from Tel Aviv University and the Open University of Israel, and has also studied at the University of Oxford.
More AI Safety companies
Official website: https://www.hirundo.io