Project FORGE is a structured 18-month pathway from programming fundamentals to industrial optimization expertise: free, open, and built for the practitioners American industry actually runs on: technicians, plant engineers, and analysts who will never sit in a graduate classroom.
American industry is being asked to reindustrialize: reshore manufacturing, secure critical mineral supply chains, and rebuild defense industrial capacity. The mathematics and software to optimize these operations already exist. What does not exist, at anywhere near the required scale, is a workforce trained to apply them.
Graduate programs in operations research and industrial AI produce a few thousand specialists a year. The industrial base needs hundreds of thousands of practitioners: people on plant floors and in control rooms who can read sensor data, build a forecast, formulate a scheduling model, and know when a prediction should not be trusted.
Project FORGE exists to close that gap from the bottom of the pipeline, where the workforce actually is.
A deliberate, competency-based progression. Each phase assumes nothing from the previous one except what it explicitly taught.
Programming from first principles: variables and memory, control flow, functions, data structures, and disciplined debugging. FORGE starts at zero by design, because that is where much of the industrial workforce starts, and because expertise built on unexamined foundations fails under production pressure.
Production-quality programming across the FORGE language set, with algorithms, complexity, version control, and database fundamentals: the working toolkit of an industrial analyst.
The mathematics that powers industrial AI: linear algebra, calculus, statistics, probability, and optimization theory, taught through plant problems such as sensor data recovery and quality prediction, not abstract exercises.
Regression, classification, neural networks, time series, and anomaly detection, implemented from scratch before frameworks are introduced, and applied to manufacturing data throughout.
Mathematical programming for real operations: production scheduling, capacity planning, network and supply chain models.
Taking models to the plant floor: data pipelines, monitoring, model maintenance, and capstone projects drawn from live industrial problem sets.
Phases marked IN DEVELOPMENT are in development, with modules published on a rolling schedule. The full curriculum map, including planned release windows, is maintained on the FORGE Framework page.
The 18-month curriculum covers seven programming languages (Python, Rust, Julia, C++, R, SQL, MATLAB), mathematical foundations (linear algebra, calculus, optimization theory), and machine learning applications specific to manufacturing environments. Each module connects theory to real industrial problems: aluminum smelter optimization, steel mill scheduling, and supply chain coordination.
Every language earns its place through an industrial use case. Python for analytics and machine learning. SQL for plant historians and production databases. MATLAB for the control engineering installed base. C++ and Rust for real-time and edge systems. Julia for scientific computing and optimization. R for statistical quality control.
No prerequisites. No tuition. No application. The only requirement is the discipline to work through the pathway in order.
Most industrial AI failures are not model failures. They are foundation failures: practitioners assembling code they cannot read, trusting outputs they cannot interrogate, and deploying systems they cannot maintain. AI coding tools have made it easier than ever to produce software without understanding. In an industrial setting, unexamined software is a liability measured in scrap, downtime, and safety incidents.
FORGE therefore teaches nothing it does not first build from the ground up. Learners implement methods from scratch before touching a framework, and every abstraction is opened at least once. It is a slower start and a far stronger finish.
Project FORGE is designed to work as workforce development infrastructure, not only as individual study. If you run a plant or an operations team and want to pilot the curriculum with your technicians or engineers (self-paced, cohort-based, or mapped to your own systems), get in touch: contact@edikan.ai. Early pilot partners help shape the applied problem sets the whole program uses.
Perspectives 01
China's critical minerals advantage was built on optimization, not geology. Closing the gap is a training problem.
Perspectives 02
Depletion happens in hours. Replenishment happens in years. The gap between those clocks is an operations problem.
Models · Industrial Capacity Notebook 01
Can federal investments close the minerals workforce gap in time? A reproducible model of the pipeline arithmetic.
Phase 3 • Mathematical Foundations
When a third of a plant's sensors fail mid-run, low-rank matrix methods can reconstruct the missing readings from the structure of the data that survives.
Phase 1 • Foundations
The conditional logic behind a famous interview exercise is the same logic that runs temperature zones, quality gates, and sampling plans on a production line.
Phase 1 • Foundations
Variables are not boxes. Understanding references, copies, and mutability is the difference between debugging by reasoning and debugging by luck.