AutoAlign Emerges from Armilla AI, Unveils Firewall Solution for LLMS

AutoAlign, a Toronto-based company spun off from its parent company Armilla AI, has unveiled a groundbreaking firewall solution aimed at enhancing the safety of large language models (LLMs). Led by interim CEO and co-founder Dan Adamson, along with CTO and co-founder Rahm Hafiz, AutoAlign is set to revolutionize the AI security landscape. Adamson, who will continue his role at Armilla AI, will be succeeded by Karthik Ramakrishnan as CEO. Hafiz will transition from his position as CTO of Armilla AI to serve full-time at AutoAlign.

Founded in 2020, Armilla AI has garnered support from renowned figures like deep learning pioneer Yoshua Bengio, Y Combinator, and Two Small Fish Ventures.With a recent infusion of $6 million CAD in seed financing, Armilla AI has solidified its position as a prominent player in the AI risk assurance sector. As the AI market expanded exponentially, Armilla AI recognized the importance of delineating its security arm from its insurance business. This realization prompted the spinoff of AutoAlign, a move aimed at concentrating efforts solely on developing cutting-edge AI security solutions. By separating its security division, Armilla AI aims to avoid potential conflicts of interest and enable both entities to flourish independently. This strategic decision underscores Armilla AI’s commitment to delivering innovative solutions that address the evolving challenges of AI deployment and mitigate associated risks effectively.

AutoAlign’s groundbreaking solution, Sidecar, heralds a new era in AI security. Tailored to interact seamlessly with large language models (LLMs), Sidecar effectively addresses a myriad of risks inherent in AI applications. From mitigating hallucinations and biases to preventing data leaks and jailbreaks, Sidecar stands as a robust defense mechanism against potential vulnerabilities. By swiftly identifying and rejecting problematic responses, Sidecar ensures the integrity and reliability of AI-generated outputs. Moreover, its capability to guide models in making corrections further reinforces the trustworthiness of AI-driven systems. With Sidecar at the helm, AutoAlign is at the forefront of fortifying AI applications against emerging threats, safeguarding the integrity of data-driven decision-making processes.

AutoAlign’s collaboration with KPMG in Canada marks a significant milestone in the responsible deployment of AI technologies. By integrating Sidecar into KPMG’s internal systems, AutoAlign demonstrates its dedication to fostering trust and transparency in AI applications. Ven Adamov’s commendation underscores the importance of AutoAlign’s commitment to responsible AI integration, aligning seamlessly with KPMG’s Global Trusted AI framework. This partnership not only enhances the security and reliability of AI-driven solutions but also reinforces AutoAlign’s position as a trusted partner in facilitating ethical and accountable AI deployment practices. Together, AutoAlign and KPMG are poised to lead the way in shaping the future of responsible AI adoption and innovation.

Indeed, Sidecar’s innovative approach to AI security, combined with AutoAlign’s strategic partnership with KPMG, positions the company at the forefront of safeguarding AI applications against emerging threats. By leveraging Sidecar’s advanced capabilities to mitigate risks such as hallucinations, biases, and data leaks, AutoAlign empowers organizations to deploy AI models responsibly and with confidence. This collaboration not only ensures the integrity and reliability of AI-generated outputs but also reinforces trust and transparency in AI-driven solutions. As the demand for secure AI continues to grow, AutoAlign’s commitment to setting new standards in AI security underscores its pivotal role in shaping the future of responsible AI deployment.