Truth Table: Logic Expression Evaluator & AI Reasoning Platform

Research & Applied AGI Scope: An Applied Science framework investigating the functional values of logical expressions to drive multi-agent orchestration, autonomous LLM decision trees, and deterministic state consensus.

Truth Table is a hybrid logic-evaluation engine and multi-agent reasoning framework designed to bridge classical symbolic logic with modern probabilistic AI models. It maps complex multi-agent interactions and LLM decision pathways to formal logical truth values, delivering deterministic state evaluation, automated decision verification, and persistent semantic state memory.

---

🚀 Architecture & Key Features

• Deterministic Logic Parser: Parses boolean and multi-valued logical expressions into verifiable truth tables and evaluation graphs.
• Multi-Agent Orchestration: Directs autonomous agent consensus using formal logical bounds rather than purely probabilistic voting.
• Semantic Vector Memory: Integrates PGVector with PostgreSQL to perform similarity searches over reasoning trees and logical assertions.
• High-Performance Native Core: Built with C++ extensions and modern Python / Django backends for computational efficiency.
• Deep Learning Integration: Leverages PyTorch and TensorFlow models to align neural representations with symbolic constraints.

---

🛠️ Tech Stack & Infrastructure Architecture

• Backend Framework: Django (Python 3.10+)
• Database & Vector Store: PostgreSQL with PGVector extension
• Cloud & Infrastructure: GCP (Google Cloud Platform) leveraging identity management efficiencies
• Authentication & Identity: Auth0 integrated with GCP IAM workflows for robust user access control
• Compute & Machine Learning: C++ (VS 2022 Workload), PyTorch, TensorFlow
• Frontend Templating: Django Templates, Tailwind CSS
• Testing & Quality Assurance: Pytest-Django, Postman, Cypress
• Containerization: Docker & Docker Compose

---

🏷️ Architectural Strategy: Interface Driven Design

This platform adopts an Interface Driven Design approach, deliberately separating core application domain logic from high-performance compute infrastructure.

Phase 1: Functional Domain Validation & Identity Baseline

The application layer communicates with models via an abstract adapter interface (AbstractLLMAdapter). This allows rapid iteration on business logic, data validation pipelines, Auth0 authentication integrations, and UX flows using lightweight inference wrappers without managing heavy GPU infrastructure prematurely.

Phase 2: Infrastructure Scale-Out & Optimization

Once domain logic and authentication pipelines stabilize, the basic inference wrapper is swapped for a production-grade Inference Layer without altering application code. Optimization focuses strictly on scaling throughput and driving down compute costs via:

• In-Fight / Continuous Batching & PagedAttention: Maximizing hardware utilization and memory efficiency.
• Quantization Pipelines (FP8/INT4): VRAM footprint reduction to enable larger models on accessible hardware.
• Dedicated Serving Engines: Compiling to fused CUDA kernels via specialized backends (vLLM, SGLang, Triton Inference Server).

🛑 Design Principle: Always code to the interface (AbstractLLMAdapter), never to the implementation. If application code requires an explicit check for whether it is talking to OpenAI or vLLM, the abstraction has leaked.

---

📊 Governance & Project Tracking

All active task tracking, bug reports, feature implementations, and sprint backlogs are managed dynamically on our GitHub Project Board following a 7-Step Business Process Optimization (BPO) methodology.

👉 Access the Project Board & Research Roadmap

---

💻 Getting Started

Prerequisites

• Docker and Docker Compose
• Python 3.10+
• PostgreSQL with PGVector extension enabled

Setup with Docker Compose

1. Clone the repository:

   git clone https://github.com/brucethagwana/truthtable.git
   cd truthtable

---

📜 License & Credits

Distributed under the MIT License. See LICENSE for full details.

• Author & Lead Researcher: Bruce Thagwana
• Core Focus: Multi-Agent Logic Orchestration & Autonomous LLM Reasoning Research

---

💬 Citation & Acknowledgments

If you use this framework or research matrix in your work, please cite it as:

@software{thagwana2026truthtable,
  author = {Thagwana, Bruce},
  title = {Truth Table: Multi-Agentic Logic & Reasoning Framework},
  year = {2026},
  publisher = {GitHub},
  journal = {GitHub repository},
  url = {https://github.com/brucethagwana/truthtable}
}

---

Special thanks to the open-source community and researchers pioneering deterministic logic evaluation in multi-agent systems.