About Schemac
Schemac was founded with a dual mission: advancing language model research while creating seamless database integration solutions for businesses. We operate at the intersection of cutting-edge AI research and practical enterprise applications.
Our team combines expertise in language model architecture, fine-tuning methodologies, and intuitive user experience design to deliver database solutions that are both powerful and accessible to business users without specialized technical knowledge.
Development
For any technical suggestions (including applications and language models), bug reports, or feature requests, please reach out to our lead developer.
Our Focus Areas
Research
We leverage the latest pretraining and fine-tuning technologies and architectures to improve performance and reduce costs for our language models. Our research focuses on optimizing models specifically for database interaction and query generation.
Integration
We develop easy-to-use interfaces and logical frameworks that make advanced AI capabilities accessible to business users. Our solutions integrate with existing database infrastructure and workflows, minimizing disruption while maximizing value.
Join Us
We are working on new features and exploring different ways to further reduce cost and improve our model's performance over data science benchmarks. Our current tech stack: React + Next.js + Postgres.
For exploring different model architectures, we are looking at exploring pathways like blockwise FP8, multihead latent attention, and parallel training techniques.
If you're interested in our work or would like to collaborate, please contact our lead developer at nan@schemac.com for more information.