Look, mah models — without containers, Databricks, or heavyweight MLOps.
GenericML is a .NET/C# framework that lets domain experts and software teams build fast, explainable machine‑learning micro‑models inside real applications (not notebooks).
Core idea: make model generation cheap and repeatable so teams can iterate to true product‑market fit.
Human‑in‑the‑Loop (HITL)
Experts validate labels, features, and predictions — boosting trust & adoption.
Data Vectors
Decision‑ready inputs: floats + float[] time‑series/distributions, type‑safe in C#.
Knowledge Graph Spine
Neo4j for provenance + multi‑hop reasoning + Graph‑RAG (graph + vector hybrid).
10 seconds version
GenericML lets domain experts and .NET teams build fast, explainable ML micro‑models inside production applications,
with HITL validation and a Neo4j knowledge graph for provenance and Graph‑RAG.
Why projects stall
- Slow iteration and brittle integration between data‑science tooling and product teams.
- Business experts can’t meaningfully test, tune, or validate models.
- Explainability, lineage, and governance are bolted on too late.
How GenericML flips it
- CPU‑only AutoML micro‑models (fast, low infra) — even on laptops.
- HITL workflows built‑in (review loops, approvals, drift triage).
- Model packs + ensemble policies + evidence bundles — everything traceable.
If you want more details
https://genericml.odoo.com/