Nexora Research produces peer-reviewed science, not press releases. Every paper represents a genuine contribution that either ships into a product or advances foundational knowledge at the frontier.
Four divisions. One mandate: push the boundary of what is technically possible.
AI Division
Parameter-efficient fine-tuning, sparse architectures, and real-time inference on constrained hardware.
Semiconductors
Sub-2nm transistor design, advanced packaging, and AI-accelerated EDA for next-generation SoCs.
Space Technology
Low-thrust trajectory optimization, ion propulsion efficiency, and LEO/MEO constellation design.
Innovation Labs
Error mitigation on NISQ devices, hybrid quantum-classical algorithms for optimization problems.
A representative sample from our research output. Full bibliography available on request.
Raj Mehta, Ananya Iyer, D. Krishnan
We introduce SparseEdge-MoE, a novel mixture-of-experts architecture that achieves 94.2% of GPT-4 benchmark accuracy while operating within a 10W thermal envelope, enabling on-device LLM inference without cloud dependency.
Priya Nair, Chen Wei, Suresh Kumar
A comprehensive simulation study of GAA-FET device physics at 1.4nm node geometries, demonstrating viable pathways for continued Moore's Law scaling through multi-bridge channel geometries and high-k dielectrics.
Vikram Sinha, James O'Brien, Priya Nair
A novel ISL relay protocol reduces end-to-end latency in low-earth-orbit constellations by 38% compared to current ground-relay architectures, enabling near real-time global internet coverage.
D. Krishnan, Ananya Iyer, R. Patel
We demonstrate a symmetry-verification technique that reduces logical error rates in NISQ devices by 3.1x without hardware overhead, advancing the timeline to fault-tolerant quantum computation.
Chen Wei, Suresh Kumar, Raj Mehta
Applying NeRF-based volumetric reconstruction to sub-10nm wafer defect imaging, achieving 99.7% detection precision with 18x faster throughput than traditional SEM inspection pipelines.
Vikram Sinha, Fei Zhang, James O'Brien
Particle-swarm-optimized formation control for constellations of 6U CubeSats enables self-healing orbital geometries, reducing station-keeping propellant usage by 44% over traditional ground-commanded maneuvers.
We are actively seeking university and industry research partnerships across all divisions. If your work intersects with ours, we want to talk.
Start a Research Partnership