We collect, curate, and create open datasets for power semiconductor reliability. We build AI models that predict device lifetime — free, reproducible, and deployable anywhere.
We find, curate, and document every open semiconductor reliability dataset in one place. NASA, university labs, published papers — all indexed, all free.
Where open data is missing, we create it. Synthetic aging profiles, Devsim simulation sweeps, and device-level characterisation — original and citable.
GBR + LSTM ensembles trained on real run-to-failure data. Predict Remaining Useful Life in milliseconds. Deploy anywhere — HuggingFace, local, or your own server.
REST APIs, web UIs, and simulation pipelines — all open source. From RISC-V CPU design to MOSFET lifetime prediction, everything ships with reproducible code.
त्रिकाल · Data · Physics · AI
Tri = three pillars. Kal = time (Sanskrit: past, present, future). Trikal predicts Remaining Useful Life of Silicon MOSFETs using a GBR + LSTM ensemble trained on NASA thermal overstress aging data.
Open-hardware CPU series built on the RISC-V RVA23 profile. From single-cycle to out-of-order execution, designed for AI, machine learning, and data center applications. Verified with ModelSim using RISC-V assembly tests.
View on GitHub ↗Open source, open data, open community.
