constraint-solving

Computational invention engine with Forge: builds software from natural-language requirements, runs it in isolation, and packages only independently validated builds.

Universal dependency resolver using SAT solvers - works for software packages, course prerequisites, books, certifications, and any system with logical dependencies

Certified frame-first SAT middleware — decide structured regions (2-SAT · GF(2) parity · counting) before CDCL, and independently verify every verdict (model replay · DRAT). A research harness for where SAT hardness lives.

Smart class timetable scheduling with constraint optimization.

An algorithm for solving the wordle.

Contradiction-native research runtime for verifier-backed ontology revision, structural abstraction invention, and replayable proof artifacts.

SC-TSE integration MVP for deterministic site-layout computation and validation.

Validated Sudoku solver with a reusable Python API, CLI, editable Tkinter board, and behavioral tests.

A semantic DSL for comic panels, compiled to SVG. Deterministic geometry and constraint solving, no generative image model. Deliberately AI-generated: a training ground for working with AI coding tools.

Unique-solution Kakuro puzzle generator with configurable difficulty, deterministic generation, SQLite persistence, multiprocessing and print-ready PDF booklet export.

Experimental offline-first Python CLI for validated specifications, context snapshots, host-execution intents, and bounded finite-domain solving; it does not call LLMs or launch agents.

Neuro-symbolic Wordoku solver with PyTorch-guided, contradiction-pruned beam search.

Computer-vision and constraint-solving research for Nonogram.com puzzle analysis on user-controlled Android devices.

🤖 Intelligent Wordle AI Bot with constraint-solving algorithm. Features smart decision-making, multi-game modes, 527-word vocabulary, and 100% test coverage. Ready for production deployment with Docker and CI/CD.

Visualize the classic N-Queens puzzle 👑 in real time using AI search algorithms — Hill Climbing, Simulated Annealing, and Local Beam Search. Built with Python, Pygame, and NumPy, it beautifully animates how AI solves complex constraint problems step by step.

Assignment-selection merge algorithms and constraint models

Generate, validate, repair, publish pipeline for AI training plans. The LLM proposes, deterministic code decides whether it ships. Resolves every constraint violation at zero LLM calls.

SAT-based Secret Santa assignment solver with family and cycle constraints.

Constraint-guided 3D scene assembly agent with typed scene graphs, connector validation, repair/rollback safety, Unity export, and MCP tooling.

Solves a riddle using a constraint solving programming backtracing approach