AI & ML interests
None defined yet.
Recent Activity
ZenteiQ
Foundation Models & Scientific Intelligence
Indigenous AI · Efficient Inference · Scientific Machine Learning
ZenteiQ develops foundation models and scientific AI technologies for research, engineering, and practical deployment. Dhīr is our flagship foundation-model family, complemented by ScireX, our scientific machine-learning library and research model ecosystem.
Dhīr — Flagship Foundation Models
Dhīr is ZenteiQ's indigenous foundation-model programme, designed to advance language intelligence while supporting efficient inference across different deployment environments.
The Dhīr collection brings together the model family and available inference formats:
Base Models (BF16 Weights)
- Dhīr 2B — dense foundation model with BF16 weights.
- Dhīr 5B-A1.7B — Mixture-of-Experts foundation model with approximately 1.7 billion active parameters, with BF16 weights.
Quantized GGUF Builds (8-bit & 4-bit)
- Dhīr 2B GGUF — quantized builds of the dense model, optimized for ultra-low-memory edge devices and CPU inference.
- Dhīr 5B-A1.7B GGUF — quantized MoE builds for faster, memory-efficient local deployment on consumer hardware.
Explore the collection for the current model roster, model cards, access conditions, supported inference formats, and downloads. The collection is maintained as the reference index rather than duplicating individual repositories here.
Explore the full Dhīr collection →
ScireX — Scientific Machine Learning
ScireX is ZenteiQ's scientific machine-learning library and research model ecosystem for partial differential equations (PDEs), neural operators, and data-driven scientific computing. The open-source library uses JAX/Flax.
The ScireX collection brings together trained models and related research across FNO, WNO, GINO, GIWO, and Transolver, covering scientific and engineering benchmarks. Visit the collection for the current model inventory and associated research references.
Explore the ScireX collection →
ScireX source code and documentation
Model Access, Licensing & Research
Some repositories may be private or access-controlled. Model availability, license terms, intended uses, and technical requirements are specified in each individual repository. The ScireX library and individual model weights may use different licenses. Papers linked in a collection are research references and do not, by themselves, imply ZenteiQ authorship.
Website: zenteiq.ai · Research enquiries: research@zenteiq.ai
ZenteiQ AiTech Innovations · Born at IISc Bangalore · Bengaluru, India