WithIn Us AI

Independent AI research organization building practical, open intelligence systems.

Focus areas

• Reasoning datasets — large-scale chain-of-thought, agent-trace, and distillation datasets curated for pretraining and fine-tuning reasoning models. Built with reproducible pipelines: global deduplication, deterministic shuffling, and regression-gated releases.
• Open models — fine-tuned, merged, and quantized (GGUF) language models with an emphasis on coding and reasoning, released for local and resource-constrained deployment.
• Training infrastructure — tooling and pipelines for dataset curation, continued pretraining, and evaluation.

On the Hub

• Models — fine-tunes, merges, and GGUF quantizations
• Datasets — distillation and reasoning datasets
• Collections — curated sets

Connect

• GitHub: WithInUsAI
• X: @WithInUsAI