Learning from Experience Lab
We are a research lab building the methods that make AI agents learn from their own experience. We improve both model weights and harness — improving performance, cutting cost, and enabling full data privacy.
Unreliable performance, rising model costs, sensitive data concerns. We solve all three.
Off-the-shelf models get 50–60% on domain tasks where you need 90%+. We close the gap with RL and in-context learning.
Frontier API costs scale linearly. Smart routing and distillation to smaller models cut your bill by up to 90%.
Data never leaves your organisation. We train and deploy models on your infrastructure with full regulatory compliance.
An agent is more than just a model. It is the model plus everything around it.
Model
weights
Harness
The model generates responses. The harness is everything around it. We improve both weights and harness continuously from real usage data.
From scoping to production — and continuous improvement after deployment.
See Our Offering & Engagement ModelOur Offering
Pick your biggest challenge. We show you exactly how we solve it.
Wrong decisions, constant escalations, inconsistent quality. They should be getting better, not staying stuck.
Your agent acts in a real or simulated environment and learns with every iteration.
No simulator? We build a training world from your data, then run RL on it.
User corrections and system feedback flow into the agent’s memory for immediate improvement without retraining.
Expensive API calls, oversized models for simple tasks. Scaling up means paying more without efficiency gains.
Automatically route easy requests to smaller models. Same quality, lower bill.
Compress frontier-model knowledge into a small model you own and run yourself.
Sensitive data, regulatory requirements, vendor lock-in. You need AI that runs in-house.
Trained models run on your servers, connected to your tools. Full data sovereignty.
Requirements
Technical Approach
How we work with you from first conversation to continuous improvement.
We start with a scoping conversation to define target tasks, success metrics, and pricing. Then we instrument your data, train your models, and deploy — with continuous improvement after launch.
Contact
What’s the AI problem you want solved?
Thank you!
We received your message and will get back to you shortly.
Or email us directly at info@aganthos.com
Operator and Contact:
Aganthos GmbH
c/o Robert
Müller & Tobias Magnus Schuster
Kirchgasse 20
6490 Andermatt, Switzerland
Represented by: Robert Müller, Chairman of the Management Board (Vorsitzender der Geschäftsführung)
Contact: info@aganthos.com · www.aganthos.com
Commercial Register (Switzerland):
Registered
Company Name: Aganthos GmbH
Commercial Register Office: Canton of Uri
Enterprise Identification
Number (UID): CHE-239.273.202
Date of Incorporation: January 13, 2026
Ownership: Wholly owned subsidiary of Aganthos, Inc. (USA).
Parent Company (USA):
Aganthos, Inc. (A Delaware
Corporation)
File Number: 10351069
Incorporated on: September 30, 2025
Registered Agent: Legalinc
Corporate Services Inc., Newark, Delaware
Liability for Content: While we strive to keep the information on this website accurate and up-to-date, Aganthos GmbH accepts no liability for the topicality, correctness, completeness, or quality of the information provided.
Liability for Links: Our website may contain links to external websites. We have no influence on the content of these external sites and therefore accept no liability for them.
Copyright: The content and works published on this website are governed by the copyright laws of Switzerland. Any duplication, processing, distribution, or any form of utilization beyond the scope of copyright law shall require the prior written consent of the author or authors in question.
Last updated: January 2026
We take the protection of your personal data very seriously. Aganthos GmbH (“we”, “us”) processes personal data in accordance with the Swiss Federal Act on Data Protection (FADP/nDSG) and, where applicable, the European General Data Protection Regulation (GDPR). This privacy policy explains how we collect and process personal data when you visit our website www.aganthos.com or contact us.
Aganthos GmbH
c/o Robert Müller & Tobias Magnus Schuster
Kirchgasse 20, 6490
Andermatt, Switzerland
Email: info@aganthos.com
Represented by: Robert Müller (Chairman)
3.1 Server Log Files: When you visit our website, the hosting provider automatically collects and stores information in server log files: browser type and version, operating system, referrer URL, hostname, time of request, and IP address. This data is not combined with other data sources.
3.2 Contact Form and Email: If you send us inquiries via email or contact forms, your details will be stored for the purpose of processing the request. We do not pass on this data without your consent.
Our company is a wholly-owned subsidiary of Aganthos, Inc. (Delaware, USA). If tools or service providers based in the USA are used, your personal data may be transferred to US servers. We ensure that data transfers are protected by appropriate safeguards, such as the Swiss-U.S. Data Privacy Framework or Standard Contractual Clauses (SCCs).
Our website may use cookies to enhance user experience. You can set your browser to refuse cookies.
We use SSL/TLS encryption for security reasons and to protect the transmission of confidential content.
Under the FADP (and GDPR where applicable), you have the right to: access, rectification, deletion, restriction of processing, and data portability. To exercise these rights, contact us at info@aganthos.com.
We reserve the right to adapt this privacy policy to comply with current legal requirements or to reflect changes to our services.