Research

Notes we are willing to stand behind.

Featured · Vortex

Internal Research Report

Vortex: Progress Report on a Compact Language-Model Research Program

A functioning research prototype and training pipeline focused on useful instruction following, school-level reasoning, and lightweight local inference from roughly 175 million trainable parameters.

By North ML12 min read

Read the record

All notes

01HorizonHorizon: Research Without the Conversational PaddingHorizon turns a question into a structured research artifact: direct findings, visible uncertainty, and evidence you can inspect without the sentimental filler.6 min
02RetrievalEvidence First: Building Answers You Can InspectRetrieval and synthesis designed to keep claims separable from conclusions.5 min
03ModelsSmall Models, Serious Research EconomicsCompact model experiments focused on capability per parameter and visible API cost.6 min
04WorkflowFrom Question to Research ArtifactA workflow that prioritizes usable output over a long conversational performance.5 min
05Model LineageAurora Proelia: Small Models as a Serious Systems ConstraintA compact model program for testing deliberate reasoning, instruction following, and deployment tradeoffs on accessible hardware.7 min
06ProductHorizon Research: Direct Answers, Inspectable ClaimsA research workspace shaped around findings, evidence, uncertainty, and next actions.5 min
07EconomicsPaper-Scale Synthesis Under a Visible Cost CeilingAn early workflow target for producing mid-sized research output while keeping inference spend legible.5 min
08MethodEvaluation Before EloquenceWhy North separates internal targets from independent benchmarks and keeps caveats attached to every claim.6 min