E-Commerce Platforms
Three full-stack business platforms with payment processing, print-on-demand fulfillment, and admin systems
Tech Stack
PHP 8.0+
MySQL
JavaScript ES6+
Stripe API
Gelato API
Printful API
PHPMailer
PHPUnit
Three production business websites with complete payment processing, dual print-on-demand API integrations, webhook systems, bilingual support (French/English), admin dashboards, and test suites.
Adinav Elegance
3 Revenue Streams: Digital downloads, art prints (Gelato API), merchandise (Printful API)
Admin Dashboard: Order management, revenue analytics, fulfillment status tracking
Test Suite: PHPUnit covering download tokens, order validation, product catalog
Webhook Systems: Stripe, Gelato, and Printful webhooks with signature verification
Adinav Consultancy
B2B Advisory Platform: Productized consulting services for startup founders with Stripe checkout
File Upload System: Multi-file upload with MIME validation and secure storage
Security: HSTS, rate limiting, honeypot bot detection, CSRF protection
Order Flow: Form → File upload → Stripe session → Webhook → Admin notification
Common Architecture
Bilingual: Full English/French with browser detection and hreflang SEO
Email System: PHPMailer SMTP with bilingual templates and order confirmations
Database: PDO prepared statements with parameterized queries
Adinav Protection: Cybersecurity services platform (detailed in separate section)
60,313
Total Lines
3
Live Sites
3
Payment APIs
1,333
Test Lines
Live verification: All platforms publicly accessible. Security ratings: 2× A+, 1× A on securityheaders.com
Cybersecurity Services Platform
Full-stack security platform with production website, professional OSINT toolkit, and case management system
Tech Stack
Python
Flask
PHP
HTML/CSS/JS
dnspython
HIBP API
PHPMailer
Three-component cybersecurity platform: bilingual production website (adinavprotection.fr), professional OSINT toolkit with 8 security research tools, and account recovery case management system.
OSINT Toolkit
Exploit Scanner: SQL injection, XSS, LFI/RFI, command injection testing
Vulnerability Scanner: Infrastructure scanning, subdomain enumeration, SSL analysis, sensitive file detection
Username Hunter: Account discovery across 51+ platforms (social, professional, developer, gaming)
Google Dorker: Google dork generation for security research with web interface
Website Security
CSRF Protection: Session-based tokens with secure expiry
Rate Limiting: IP-based request throttling
Security Headers: HSTS, X-Frame-Options, scanner detection
Email Security: SMTPS encryption with secure credential handling
Account Recovery Tools
Case Management: Workflow states from intake to resolution
Breach Checker: HIBP API integration with privacy-preserving k-anonymity
6 Platform Playbooks: Instagram, Facebook, WhatsApp, Google/Gmail, Twitter/X, LinkedIn
Service Documentation: Playbooks, templates, and operational guides
12,354
Lines of Code
8
Security Tools
51+
Platforms Scanned
6
Recovery Playbooks
Live verification: Website live with A+ security rating. OSINT toolkit tested on authorized targets.
Digital Agency Platform
Premium bilingual website for web development services with outcome-based pricing
Tech Stack
Astro 5.x
Tailwind CSS v4
PHP 8.0+
GSAP
PHPMailer
Full-featured digital agency website built with Astro for static site generation. Features 29 bilingual pages (English/French), premium visual design with animations, and secure contact form backend.
Key Features
29 Bilingual Pages: Complete English/French with route-based i18n and hreflang SEO
Premium Design: Gradient backgrounds, glass effects, animated hero, responsive breakpoints
5 Service Tiers: Outcome-based pricing from Online Presence to Custom Solutions
Contact Backend: PHP with CSRF protection, rate limiting, honeypot spam detection
Security Headers: A+ rating target with HSTS, CSP, X-Frame-Options, scanner blocking
SEO Optimized: JSON-LD schemas, sitemap, semantic HTML, lazy loading images
16,043
Lines of Code
29
Bilingual Pages
8
Components
25+
Sessions
Live verification: Website publicly accessible with A+ security headers target
Personal AI System
A multi-agent cognitive architecture for personal and family use
Tech Stack
Python 3.13
FastAPI
PostgreSQL
pgvector
Claude API
Ollama
Whisper
Piper TTS
A production-grade personal AI system featuring 5 specialized agents (Memory, Planner, Analyst, Critic, Workspace), 3-layer memory architecture, and family-safe web interface with role-based access control.
Key Features
5-Agent Architecture: Specialized agents for memory, planning, analysis, critique, and file system access
4-Tier LLM Routing: Cost optimization from local Ollama (free) to Claude API (complex tasks)
3-Layer Memory: Short-term context, semantic search (pgvector), and structured PostgreSQL storage
Family Safety: Role-based profiles with age-appropriate content filtering
Voice Interface: Speech-to-text (Whisper) and text-to-speech (Piper) for hands-free use
Autonomous Tasks: Job system with approval gates for multi-step project execution
Architecture
User Input → Orchestrator (values enforcement)
↓
Memory Agent (load context from 3 layers)
↓
Planner Agent (task decomposition)
↓
Analyst Agent (main reasoning)
↓
Critic Agent (quality + safety check)
↓
Response → Memory storage
20,000+
Lines of Code
164
Passing Tests
5
AI Agents
44
Evolution Logs
Documentation available: Full architecture docs, test results, evolution logs tracking 44 development sessions
BookPulse - Amazon Book Intelligence
Full-stack dashboard for tracking Amazon bestseller rankings with AI-generated market briefings
Tech Stack
Next.js 16
TypeScript
Prisma
PostgreSQL
Supabase
Tailwind CSS
React Query
Recharts
Claude API
Full-stack Next.js application tracking Amazon book rankings across categories. Features 7 API routes, 11 React components, 30-day historical charts, and AI integration for generating market analysis briefings.
Key Features
7 API Routes: Books, categories, movers, briefings, analyze — with filtering, sorting, pagination
30-Day Charts: Rank history, price trends, review counts per book using Recharts
AI Briefings: LLM generates daily/weekly/trend/gap analysis on demand
6 Custom Hooks: useBooks, useMovers, useBriefings, useDashboardStats with React Query
Responsive Dashboard: Stats cards, movers widget, book grids — mobile to desktop
Database: PostgreSQL with 4 tables tracking books and historical snapshots
Architecture
Next.js App Router (7 pages)
├── /api/books, /api/movers, /api/briefings, /api/analyze
├── PostgreSQL (Supabase) with Prisma ORM
├── React Query for client state
└── LLM API for AI briefing generation
3,540
Lines of Code
11
Components
7
API Routes
4,500
Data Records
Live verification: Deployed on Vercel with Supabase database
AI Avatar Pipeline
Complete AI generation pipeline: image, video, and audio synthesis
Tech Stack
ComfyUI
Stable Diffusion XL
PuLID
EVA-CLIP
LivePortrait
Edge-TTS
InsightFace
Three-pillar AI content generation system: text-to-image with face consistency, talking head video animation, and voice synthesis. 4 production workflows with iterative development history.
Image Generation
Text-to-Image: Photorealistic portraits using Juggernaut XL at high resolution
Face Consistency: PuLID + EVA-CLIP embeddings for identity preservation across generations
Full-Body Generation: Complete character generation with face detail post-processing
Iterative Development: Multiple workflow versions tested and refined
Video & Audio
Talking Head Animation: Portrait + driving video produces MP4 with lip sync
Video Pipeline: Face alignment → Animation processing → Final composition
Text-to-Speech: Multiple voice options with real-time synthesis
Voice Cloning: Advanced TTS model for custom voice generation
8,700+
Lines Workflow Code
4
Production Workflows
20+
Workflow Iterations
17
Documentation Files
Documentation available: Research docs including cross-validation with multiple AI models
Trading Analysis Platform
Portfolio of 6 interconnected trading systems with 4-engine signal generation and Bloomberg-quality dashboards
Tech Stack
Python 3.10+
FastAPI
Streamlit
PostgreSQL
Celery
Redis
TA-Lib
Qiskit
Six interconnected trading projects progressing from experimental to production-ready. Features 4-engine multi-factor signal generation, 23+ REST API endpoints, and parallel market scanning (50-200 stocks in 30-90 seconds).
4-Engine Signal Architecture
Technical Engine (40%): 29+ indicators (RSI, MACD, Bollinger), pattern detection (head & shoulders, breakouts)
Fundamental Engine (30%): P/E, PEG, ROE, ROA, growth metrics, earnings analysis
Sentiment Engine (20%): SEC insider filings, NewsAPI, Reddit (praw), VADER sentiment scoring
Volume Engine (10%): VWAP, smart money detection, accumulation/distribution, volume clusters
Project Progression
Professional Trading Platform: Enterprise-grade with pattern detection and WebSocket support
AI Trading Intelligence Pro V1: Extended strategies, penny stock analysis, enhanced data providers
AI Trading Intelligence Pro: Bloomberg-style UI, parallel scanning, master signal generator
Foundation Projects: REST API foundation, broker connector, quantum simulation prototype
67,600+
Lines of Code
57
Test Files
23+
API Endpoints
150+
Doc Files
Status: Production-ready, paused due to real-time data subscription costs. Technical completion 100%.
Sports Prediction Engine (Drishti)
Production-grade football prediction engine for sports betting
Tech Stack
Python 3.11+
FastAPI
PostgreSQL
asyncpg
NumPy/SciPy
BeautifulSoup
Production-grade football prediction engine combining Poisson models with Elo ratings, form adjustments, and contrarian analysis for sports betting markets.
Prediction Models
Poisson Model: Dixon-Coles adjusted with league-specific home advantage and form multipliers
Elo Ratings: 150+ international teams across all confederations
Daily Scanner: Portfolio builder with tiered confidence levels
Grid Generator: Multiple betting strategies (Banker, Value Hunter, Contrarian)
10,500+
Lines of Code
6,400+
Historical Matches
500+
Teams Tracked
25+
CLI Scripts
Live validation: Tested with real betting to validate predictions against actual outcomes
Pattern Detection Research
Research investigating signal detection in high-entropy systems using ML and multi-AI validation
Tech Stack
Python
XGBoost
PyTorch
Qiskit
LLM APIs
PostgreSQL
Monte Carlo
Comprehensive research applying ML ensemble models, statistical analysis, and multi-AI validation to pattern detection in high-entropy data. Multiple independent approaches reached the same conclusion: validated true randomness.
ML Ensemble Approach
5-Model Ensemble: XGBoost, Statistical, QAOA Quantum, QAE Quantum, Time-Series Transformer
207+ Features: Temporal, statistical, pattern, pairwise co-occurrence
Walk-Forward Validation: Demonstrated backtest vs real-world divergence — proved overfitting risk
Key Finding: Deep learning underperformed simple models on weak signals
Statistical Analysis
20+ Statistical Tests: Chi², KS, runs, serial correlation, PRNG fingerprinting
A/B Testing: Compared AI predictions vs random baseline
Conclusion: Statistical significance ≠ predictive value
Monte Carlo: 7,000+ simulations for probability validation
Multi-AI Validation
Parallel Development: Multiple LLMs built predictors independently
Cross-Validation: Independent systems tested jointly
Ensemble Methods: Multiple predictor versions with weighted blending
Convergent Results: All AI systems reached same conclusion independently
33,000+
Lines of Code
12+
ML Models
20+
Statistical Tests
27
Evolution Logs
Scientific value: Multiple independent AI systems proved randomness — validates rigorous methodology
Quantum Market Simulation
Hybrid classical-quantum scenario forecasting with full Qiskit implementation
Tech Stack
Qiskit 1.2+
Qiskit-Aer
qiskit_algorithms
hmmlearn
Streamlit
Plotly
Full quantum-classical hybrid system for strategic scenario forecasting. Features complete Qiskit implementations of QAOA and QAE algorithms with Streamlit dashboard, integrated into the Trading Platform ecosystem.
Quantum Implementations
QAE Tail Risk: IterativeAmplitudeEstimation for VaR/CVaR calculation with state preparation and tail oracle
QAOA Regimes: Market regime clustering via QUBO with COBYLA optimizer and modularity maximization
Classical Baseline: HMM regime detection and Monte Carlo simulation for honest comparison
Visualization: Fan charts, regime visualization, and interactive Streamlit dashboard
3,900+
Lines of Code
2
Quantum Algorithms
3
Classical Baselines
1
Dashboard
Status: Fully implemented within Trading Platform. No quantum advantage claims — transparent classical baseline comparisons throughout.
Published Books
Technology strategy and AI ethics for general audiences
The AI Parenting Framework (Jan 2026)
AI guidance for parents
Kindle + Paperback
AI guidance for parents
Kindle + Paperback
AI Inequality: The Hidden Crisis of Digital Democracy (Sept 2025)
Analysis of AI's societal, economic, and democratic impacts
Kindle + Paperback
Analysis of AI's societal, economic, and democratic impacts
Kindle + Paperback
Free AI for Social Media (Sept 2025)
The Complete Growth Hacking Guide
Kindle + Paperback
The Complete Growth Hacking Guide
Kindle + Paperback
Verification: Click links above to view on Amazon