foretodata://lab Represent → Simulate → Decide → Learn
$ whoami

James Young

I build decision systems for complex domains. Across decision systems, interactive AI, and scientific machine learning, the same loop turns many interacting variables into a feasible next action. Language models can make that system conversational while reliable calculations, evidence, privacy, and approval boundaries remain explicit.

// REPRESENT

Make the outcome, choices, context, and limits explicit.

// SIMULATE

Compare realistic actions before committing resources.

// DECIDE → LEARN

Choose what can be carried out, then use the result to improve the next decision.

$ explore ./systems --group=domain

Systems in practice

./operational-decision-systems

Optimizing resources under real-world constraints

A reusable decision system for representing what drives an outcome, simulating possible allocations, choosing the best feasible plan, and learning from what happens next.

open system →
./governed-interactive-analytics

Conversational analytics with answers you can verify

A governed analytics workbench where an LLM interprets the question, tested analytical engines produce the answer, and every result includes its evidence.

open system →
./private-local-intelligence

Private AI for sensitive, long documents

An offline desktop assistant that keeps documents and model inference on the user’s device while breaking long material into a structured, reviewable analysis.

open system →
./adaptive-analytics-platform

Analytics that adapt to each organization’s data

A framework that keeps the validated analysis stable while using an LLM to understand each new data environment and create the adapters needed to run it.

open system →
./oxic-nitrogen-fixation

Choosing which genes to test for oxygen-tolerant nitrogen fixation

A decision system using scientific machine learning to combine biological evidence and rank gene candidates for scarce experimental time and resources.

open system →
./rubisco-active-learning

Choosing which protein variants to test next

A decision system using scientific machine learning to balance promising protein candidates with experiments that will reduce uncertainty.

open system →