How an 18-month open curriculum is engineered to turn industrial practitioners into optimization-capable engineers, and how the model scales beyond one website.
FORGE assumes no prior programming, no calculus, and no statistics. This is not a simplification; it is the design constraint that makes the program useful. The people American industry most needs to upskill (technicians, operators, plant engineers) are systematically excluded by programs that assume a computer science degree on day one.
Every method is implemented from first principles before any library or AI tool is introduced. A practitioner who has built gradient descent by hand can debug a model that misbehaves in production. A practitioner who has only called an API cannot.
No module ends at the whiteboard. Linear algebra is taught through sensor-network data recovery. Regression is taught through roll-force prediction. Scheduling theory is taught through steel mill and smelter operations. The industrial application is the pedagogy, not the appendix.
Each phase defines what a learner can do on exit, stated as capabilities an employer can verify, not certificates of attendance.
All core materials are free and public. Scarcity of training is the problem FORGE exists to solve; it cannot also be the business model.
| Phase | Focus | Exit competencies (abridged) | Status |
|---|---|---|---|
| 1. Foundations | Programming from first principles: memory, control flow, functions, data structures, debugging | Read, write, and debug non-trivial programs unassisted; explain what code does at the memory level | LIVE |
| 2. Core Competence | Production-quality programming across the language set; algorithms and complexity; Git; SQL fundamentals | Ship maintainable code; query and manage industrial databases; reason about performance | IN DEVELOPMENT target window TBD |
| 3. Mathematical Foundations | Linear algebra, calculus, statistics, probability, optimization theory | Formulate industrial questions mathematically; apply SVD/PCA, estimation, and basic optimization to plant data | PARTIAL first modules live |
| 4. ML for Manufacturing | Regression, classification, neural networks, time series, anomaly detection: from scratch, then frameworks | Build, validate, and stress-test predictive models on manufacturing data | IN DEVELOPMENT target window TBD |
| 5. Industrial Optimization | Linear, integer, and network programming; scheduling; capacity and supply chain models | Formulate and solve production scheduling and supply chain problems; interpret duals and sensitivity | IN DEVELOPMENT target window TBD |
| 6. Deployment & Capstones | Data pipelines, monitoring, model maintenance, capstone projects | Take a model from notebook to monitored production use; complete an end-to-end industrial capstone | IN DEVELOPMENT target window TBD |
Seven languages is a deliberate curriculum decision, not breadth for its own sake. Industrial environments are heterogeneous, and each language maps to a layer of the real stack:
Learners achieve working fluency across the set and depth in the layers their role demands.
FORGE's worked examples are drawn from operations of strategic significance to American industry, so that training capacity and national industrial priorities compound rather than compete:
On the roadmap: Optimization for Critical Supply Chains & Defense Industrial Capacity. A dedicated advanced track applying the full FORGE toolkit to publicly documented problems in industrial surge capacity: munitions production scheduling, strategic material inventory modeling, and processing bottleneck analysis, built exclusively on open data and published sources. Module outline to be published on this page.
FORGE is engineered to grow through three deliberate stages:
Self-paced curriculum, public repository, open problem sets. Success measure: complete, coherent coverage of Phases 1–4.
Facilitated cohorts with fixed timelines, peer review, and instructor office hours: the structure most adult learners need to finish what open courseware starts.
The curriculum packaged for manufacturers (in-house upskilling tracks mapped to their own systems) and for community colleges and training programs (technician-level pathways feeding the same progression). This is where a website becomes workforce infrastructure.
Manufacturers interested in Stage 2–3 pilots: contact@edikan.ai.
The edikan.ai platform launched in September 2025 as an open learning site in industrial AI. In August 2026, the program was restructured as Project FORGE: the curriculum architecture published on this page was formalized, early modules were revised into the current program voice, and the Perspectives series was launched. The analytical work behind the program, spanning industrial optimization, critical mineral supply chains, and workforce capacity, has been in development since 2024 through the founder's doctoral research and professional practice. Original publish dates are preserved on revised modules, and all dates on this site reflect actual publication events.