top of page

The Future of Autonomy is Squint AI

Helping technology reflect on decisions


Over-confidence is bad in humans, it is worse in AI models. Squint AI helps build trust in autonomous decision making.

Even the best AI models are capable of high-confidence mistakes. As humans we squint when we are unsure of what we are looking at. It's a second look; a confidence builder. Squint AI provides a platform of algorithms that enable AI models to Squint when the data is just not good enough


We work with leading edge partners around the world, providing access to game-changing AI solutions

Understand your models

Squint AI empowers AI application engineers to understand how well their models are representing the training data. Later, at production time, the Squint AI platform evaluates how well the trained models perform, and how close the training data matches the real-world data seen in production.

Business meeting
Brain Scan

Imagine a CatScan for your models

Squint AI's patented technology is a runtime CatScan for autonomous software. Squint AI is a platform for analyzing decision making at runtime by inspecting the internal states of the active AI model at the time of a decision. The analysis is used to calculate the likelihood of a mistake in a given decision.


A different approach to explaining AI

From automotive, to industrial robotics and medical devices, when autonomy is involved, explainability is required. Squint brings explainable AI to the runtime.


We’re looking for innovative talent to join our team. See all positions and submit your CV.

LLM Applications Engineer

We are looking for an exceptional Machine Learning applications engineer with experience aligning Microsoft Copilot and/or OpenAI's ChatGPT models to downstream tasks using prompt-engineering, chain-of-thought prompting, and fine-tuning.

bottom of page