A
Azul AI
End-to-end fine-tuning & inference platform
Private beta

Private Local AI,
built end to end

You have the hardware. We have the platform. Fine-tune, evaluate, and serve your own LLMs on your own GPUs. Datasets, training, RAG, and inference in one unified platform. Your data and your models never leave your machines.

SFT + RAFT + DPO · LoRA / QLoRA · RAG over pgvector

AI model visualization

The Platform

Not a wrapper around an API. Purpose-built for private, local LLM development.

Train models bigger than your GPU

A deterministic placement planner pools VRAM across mismatched GPUs, so 7B–14B LoRA and QLoRA jobs fit on the hardware you already own. No NVLink required.

The full fine-tuning toolkit

Supervised fine-tuning and Direct Preference Optimization with LoRA / QLoRA adapters and GPU-aware quantization, with every batch's loss retained uncapped for post-run analysis.

Runs that survive the night

Automatic checkpointing, heartbeats, and exact-state resume mean a multi-hour run picks up right where it left off instead of restarting from epoch one.

Know your model actually got better

Held-out probe suites graded by exact, numeric, and LLM-as-judge methods with median-of-3 sampling to cut judge variance, then charted as a pass-rate curve across a training campaign's stages.

Serve it the moment it's trained

Fine-tuned and base models served over OpenAI-compatible vLLM endpoints, with a control-plane arbiter that keeps training and inference from ever colliding on the same GPU.

Ground models in your own documents

Knowledge bases ingest your documents, chunk them, and store embeddings in Postgres pgvector for retrieval-augmented generation alongside your fine-tuned models.

On-prem
local deployment
You own the weights: keep them on your hardware, served from your own vLLM endpoint
A/B compare
inference comparison
Fine-tuned vs base completions on the same prompt, side by side
Synthetic data
dataset generation
Generate persona and custom SFT / DPO training sets, no hand-labeling
Hugging Face
model export
Push fine-tuned checkpoints straight to the Hub

Real runs, real numbers

Live figures from the training and evaluation datasets behind an actual fine-tuning campaign run on the platform. No mockups.

Held-out evaluation suite
180 contamination-free probes, stratified across 12 capabilities, graded by an LLM-as-judge with median-of-3 sampling.
DPO preference pairs by axis
573 chosen/rejected pairs, voice-balanced, targeting eight alignment behaviors.
  • Voice42.9%
  • Grounding25.3%
  • Reconciliation8.7%
  • Identity8%
  • Answerability4.9%
  • Privacy3.8%
  • Embellishment3.5%
  • Abstention2.8%
Adversarial stress suite
108 harder probes across seven attack categories: multi-hop inference, contradiction defense, and attribution boundaries.

Cloud convenience. On-prem privacy.

The platform runs anywhere, no GPU required. Training and inference run on your own workstation or datacenter. They communicate only over a queue and HTTP. This means your hardware can live in a homelab, behind a tunnel, or in your own VPC.

Control plane

No GPU
  • Web console
    Dashboard, training campaigns, eval grader, playground
  • Platform API
    Auth, multi-tenant projects, job orchestration
  • Inference arbiter
    Keeps training and live inference from fighting over a GPU
HTTP

Data plane

GPU workstation
  • Training agent
    Listens for jobs, runs the training pipeline
  • vLLM servers
    OpenAI-compatible serving, fine-tuned + base
  • Container manager
    Swaps served checkpoints, restarts vLLM

Shared backbone

PostgreSQL + pgvector
State, metrics, and vector embeddings
Runner protocol
One job manifest + callback contract for every GPU box: yours, the platform's, or HF cloud
AWS S3
Datasets & model artifacts over presigned URLs

Start with one workstation and enroll more GPU boxes as you grow. Azul AI can orchestrate and drive the whole fleet.