Custom Software
Purpose-built systems for operational problems that off-the-shelf software does not solve well.
Logarithm builds custom software, AI, optimization, and analytics around costly operational problems.
Discuss a problemTechnology chosen for the problem, not the other way around.
Purpose-built systems for operational problems that off-the-shelf software does not solve well.
Applied models and intelligent workflows where prediction, classification, or automation creates measurable value.
Scheduling, allocation, and decision systems designed to reduce time, cost, and operational friction.
Data pipelines and decision tools that convert scattered information into clear operational insight.
Start with the economics of the problem. Then engineer the solution.
Find the expensive, repetitive, or difficult process.
Measure its cost, frequency, constraints, and impact.
Create the smallest system capable of solving it.
Measure the result against agreed success criteria.
Early systems focused on labor-intensive operations, data, and decision-making.
Led development of an AI/ML backbone for improving patient triage and subspecialty referral within a major health system.
↗Developed a YOLO-based computer vision system and supporting meta-models that reduced counting error and error variance by more than 60% for a leading US poultry company.
↗Developed symbolic regression and grammatical evolution frameworks for efficiently searching complex model spaces and discovering predictive structures.
↗Built a genetic programming pipeline that automatically generates, optimizes, validates, and combines predictive models for financial market behavior.
↗Designed a multi-layer neural network architecture that distributes learning across feature and time dimensions before aggregating predictions into a unified meta-model.
↗Start with the problem. We can determine whether software should solve it.
logan@logxco.com