The capital injection includes participation from Accel, Boldcap, Nexus Venture Partners, Premji Invest, and Unbound. CEO Ranjan Rajagopalan argues that while LLMs provide the flexibility to navigate natural language queries, they remain prone to errors unacceptable in sensitive sectors. By layering a deterministic verification system over the model, Pramaana forces AI outputs to align with codified rules rather than probabilistic guessing.
In section Startups & Technology
Pramaana Labs secures $27M to bring mathematical rigor to AI
Pramaana Labs has secured $27 million in seed funding led by Khosla Ventures to inject formal verification into enterprise AI. The startup aims to eliminate hallucinations in high-stakes fields like tax law and drug discovery by tethering large language models to the deterministic precision of the LEAN programming language.
Drawing inspiration from France’s CATALA project, which successfully translated tax codes into executable logic, the company is building bespoke verification systems for each vertical. Domain experts are central to this architecture: former IRS commissioner Danny Werfel is advising on tax law applications, while faculty from UC Berkeley, IIT Delhi, and IIT Madras provide oversight for cybersecurity and pharmaceutical research. Rajagopalan views the current limitations of AI as a failure of formalization, suggesting that once complex regulatory or scientific frameworks are properly codified, the reasoning process shifts from speculative to deterministic.
Comments (0)
No comments yet. Be the first!