Mitrasish Mukherjee
Co-founder & CTO, Spheron
Co-founder & CTO at Spheron. Writes about GPU clouds, AI infrastructure, and what it actually takes to run large-scale model training and inference.
Mitrasish is Co-Founder & CTO of Spheron, an enterprise GPU rental marketplace that aggregates capacity from vetted data center partners worldwide. He works closely with AI teams shipping LLM training, fine-tuning, and inference at scale, and writes most of the engineering and product content on the Spheron blog. Before Spheron he spent years building developer infrastructure across compute, storage, and orchestration.
Posts by Mitrasish

Cost Per Million Tokens: How to Calculate Your Inference Bill
Sep 3, 2026
GPU Spec Sheet vs Real World Performance Explained (2026)
Sep 3, 2026
Speculative Decoding Explained: Skip the GPU Upgrade
Sep 3, 2026
FP8 vs BF16: What You Actually Lose in Accuracy (2026)
Sep 2, 2026
Tensor Parallelism vs Pipeline Parallelism: Splitting a Model Across GPUs
Sep 2, 2026
Continuous Batching vLLM Explained: Cheaper LLM Inference
Sep 1, 2026
FlashAttention Explained: How It Cuts LLM Inference Costs
Aug 31, 2026
HipKittens AMD Kernels: MI355X Performance in 2026
Aug 31, 2026
CoreWeave vs Crusoe: A GPU Cloud Pricing Comparison 2026 for H100 and B200
Aug 30, 2026Try It on Real GPUs
The GPUs behind these guides are the ones you can rent here: H100s, H200s, B200s, and more, billed per minute with no contracts and no minimum. Pick one and you are live in under two minutes.