slipa beta

§ 01Platform

Spot GPU orchestration
honed for async AI workloads.

slipa routes checkpoint-friendly jobs across five spot providers, recovers from evictions, and migrates mid-run when a cheaper GPU opens up. 30–70% below on-demand.

Request access → See the flow

§ 02How it works

Four commands. A running GPU.

Describe the workload, set a budget. slipa picks the cheapest live spot GPU, provisions it, runs your container, handles eviction recovery, and hands back the result.

# install + connect $ pip install slipa $ slipa configure # boot a GPU, prove CUDA, mark complete. ~$0.05. $ slipa run hello-gpu --duration 60 # fine-tune an open LLM with a per-run budget ceiling. $ slipa run finetune \ --model meta-llama/Llama-3.1-8B \ --dataset tatsu-lab/alpaca \ --method qlora \ --max-budget 5.00 # follow, download. $ slipa logs <id> $ slipa download <id>

§ 03Workloads

One router. Many workload shapes.

finetune
LoRA / QLoRA / full fine-tune of an open-source LLM. Axolotl under the hood.
Live
hello-gpu
Scaffolding. Boots a GPU, runs a CUDA sanity check, ticks, completes.
Live
batch-infer
Synthetic data, embeddings, bulk classification, offline scoring.
Next
rl-rollouts
Stateless trajectory generation for post-training loops.
Soon
agent
Long-running checkpoint-friendly agent tasks.
Soon
Routes across five spot providers · finds the cheapest · survives evictions

§ 04Access

Request an invite.

Private beta. Leave an address and we'll reach out when capacity opens for your workload shape.

No commitment. No newsletter.