Sarvam reposted this
Open source AI models ensure that more people around the world have access to the benefits and opportunities of AI, that power isn’t concentrated in the hands of a small number of companies, and that the technology can be deployed more evenly and safely across society. Developers know better than anyone the value of an open approach to AI, and, earlier this month at the Global Open Source Innovation Summit, we were proud to host 70 Llama developers from 22 countries at our Menlo Park headquarters – this is such an inspiring group! During two days of panels, demos, and discussions, developers showed us how they’re making a positive impact with Llama. Here are just a few examples: Brandon Mitchell, and Hanzla Ramey at WriteSea, a career services startup in Oklahoma, used Llama to build Job Search Genius, a tool that helps candidates write their resume, practice interviews, and learn salary negotiation. Tetsuro Ariyoshi from Quixotiks, a startup in Japan, uses Llama 3.2 to enable cognitive psychology and other support that spans dementia prevention, elderly care, mental health, healthcare behaviors, and coaching. Yasmin Abdi from noHack, a startup in Virginia, uses Llama to help companies proactively identify, mitigate, and manage insider risks by reducing risk identification time by 30%. Claus Lang from Kodex AI, a startup in Germany, used Llama to build the GenAI Compliance Coworker, which fully automates the tasks of a regulatory analyst and achieves significant time savings for the financial industry. Dunstan Matekenya, a data scientist working in collaboration with Flowinformatics, a Malawi-based firm, used Llama to build a chatbot capable of conversing in English and Chichewa to answer questions based on databases and textual information, addressing challenges like food insecurity. Edgar Huaranga from Hiperderecho, a nonprofit in Peru, used Llama to build aLex, the Augmented Legal Experience, which democratizes access to and understanding of the Peruvian legal system. Tushar Goswamy represented Sarvam, the Indian startup which used Llama to develop Shuka v1, India’s first Open-source Audio language model, which is proficient in 10 Indian languages. Bastien Bernath, Etienne Boisson, and Xavier Theimer-Lienhard from the Lab for Intelligent Global Health & Humanitarian Response Technologies at Yale School of Medicine and EPFL used Llama to build Meditron – the world’s best-performing open-source LLM designed for medical guidance – and to run their Massive Online Open Validation and Evaluation (MOOVE). Sakura Seet and Lee Jian Wei Darren from Perfingo, a startup in Singapore, used Llama 3.1 to build an AI companion that delivers financial guidance through personalized, psychologically-aware advice. Iyan Sonesra, Noor Ali, and William Zhang from Fynopsis, a startup in Texas – and winners of the recent Austin Llama Hackathon – use Llama to support document management, processing, and sharing in virtual data room (VDR) services.
-
-
-
-
-
+14