Work

Selected projects spanning AI tooling, e-commerce, and autonomous systems.

Drew Garman builds AI-powered software and websites. Key projects include DBNT, an open-source Python protocol for AI feedback loops (available on PyPI), and Borussia Minerals, a live e-commerce site for premium mineral specimens.

Do Better Next Time

live

Universal learning protocol for AI feedback loops

Challenge

AI agents repeat the same mistakes across sessions. Traditional error logging captures what failed but not why, and nothing teaches the system to improve. Success patterns are ignored entirely.

Solution

Open-source Python protocol that encodes both success and failure patterns with weighted signals (success 1.5x over failure), FSRS spaced-repetition decay for rule relevance, and pluggable adapters for any AI framework.

Result

84 tests, zero dependencies beyond click, pip-installable. FSRS decay keeps rules fresh — unused learnings fade naturally. Ships with Claude Code hooks, CLI, and plugin system. Published on PyPI.

PythonFSRSCLIOpen Sourcepip

Borussia Minerals

live

Client: Boris Dimov

Professional e-commerce for premium mineral specimens

Challenge

Needed a professional web presence for a premium mineral specimen business showcasing Arizona's Fat Jack Mine specimens.

Solution

Designed and built a clean, museum-quality e-commerce site with responsive design and contact integration.

Result

Live professional presence matching the premium nature of the product. Phase 2 in progress: Stripe payments, 360° product photography, order management.

Next.jsReactTailwind CSSTypeScriptVercel

AI Newsroom

active

Voice-driven news headlines powered by AI

Challenge

Build an automated news delivery system with natural voice synthesis and editorial curation.

Solution

AI-powered news aggregation pipeline with voice synthesis delivering headlines in the style of a seasoned anchor.

Result

Automated daily news broadcasts with AI-generated voice delivery and curated headline selection.

PythonAI Voice SynthesisNLPAutomation

AI Agent Infrastructure

active

Distributed compute for autonomous AI workloads

Challenge

Build infrastructure for running AI agents across multiple machines.

Solution

Multi-node orchestration layer for local AI model serving, task routing, and autonomous execution.

Result

Private AI compute mesh running local models with cross-node coordination.

PythonTypeScriptLocal LLMsOrchestration

Trading Signal Engine

active

Autonomous market intelligence pipeline

Challenge

Automated system for processing market data and generating insights.

Solution

NLP-driven pipeline for sentiment scoring and signal generation from multiple data sources.

Result

Autonomous analysis pipeline with multi-factor scoring.

PythonNLPAutonomous Agents