MAZEL
R&D Lab

What We're Building

Internal R&D projects that push the boundaries of what we deliver for clients. Every system here was designed, built, and operated by our team — and reflects the same engineering rigor we bring to every engagement.

Full-Stack Development

Shopping Intelligence Platform

A custom-built web application that aggregates real-time pricing data across major retailers, delivering instant product comparisons and market intelligence. The platform combines a high-performance API backend with an automated web scraping engine capable of extracting live pricing from dynamic, JavaScript-rendered storefronts — and a responsive frontend that surfaces actionable insights in seconds.

Key Capabilities

  • Real-time price comparison across multiple retailers
  • Automated data extraction from dynamic web pages
  • Structured API layer with intelligent caching and rate limiting
  • Interactive frontend with live search and filtering

Demonstrates

Full-Stack App DevelopmentAPI DesignReal-Time Data Processing
FastifyPlaywrightSurrealDBReactNode.js
Network Security

Enterprise Security Architecture

A novel network security architecture that combines Palo Alto next-generation firewalls with open-source threat detection platforms to deliver enterprise-grade visibility at a fraction of traditional cost. The design leverages hardware TLS decryption and application identification paired with Suricata and Zeek for deep packet inspection and behavioral analysis — a hybrid approach with no published implementation guides at the time of design.

Key Capabilities

  • Full TLS decryption with hardware-accelerated inspection
  • Application-level identification and policy enforcement
  • Open-source IDS/IPS integration for advanced threat detection
  • High-availability failover with redundant firewall pairs

Demonstrates

Security ArchitectureCost OptimizationVendor-Agnostic Design
Palo Alto NGFWSecurityOnionSuricataZeekMikroTik
AI / ML

AI Agent Orchestration Platform

A multi-model AI agent system that autonomously triages, routes, and executes tasks across distributed compute nodes. An intelligent gating pipeline filters and escalates work based on complexity — routing simple tasks to lightweight models and reserving high-capability models for problems that demand deeper reasoning. The result is an autonomous operations layer that runs 24/7 with minimal human oversight.

Key Capabilities

  • Multi-model routing with complexity-based escalation
  • Autonomous task execution across distributed nodes
  • Quota management and intelligent resource allocation
  • Real-time relay messaging between agents and operators

Demonstrates

AI/ML IntegrationAutomationDistributed Systems
Claude APIOllamaSurrealDBTypeScriptsystemd
DevOps

Infrastructure Monitoring & Automation

A distributed monitoring and automation platform that maintains continuous visibility across multiple compute nodes, databases, and services. Automated health checks, database replication, backup pipelines, and service watchdogs ensure uptime and data integrity without manual intervention — with anomaly detection that escalates issues before they become outages.

Key Capabilities

  • Multi-node health monitoring with automated alerting
  • Database replication and bidirectional sync across sites
  • Scheduled backup pipelines with retention management
  • Service watchdogs with automatic restart and escalation

Demonstrates

DevOpsInfrastructure as CodeReliability Engineering
BashsystemdSurrealDBcronTailscale
Data Architecture

Knowledge Management System

A graph-based knowledge management platform that captures, connects, and retrieves institutional knowledge across sessions and contexts. The system models observations, claims, and procedures as a structured knowledge graph — with built-in coherence checking, claim maturity tracking, and automated procedure extraction that turns operational experience into reusable intelligence.

Key Capabilities

  • Graph-based knowledge modeling with typed relationships
  • Automated coherence checking and conflict detection
  • Claim lifecycle management with evidence linking
  • Procedural memory extraction from operational patterns

Demonstrates

Database ArchitectureData ModelingIntelligent Systems
SurrealDBTypeScriptMCP ProtocolGraph Queries

Why We Build Internally

Every platform listed here started as a real problem we needed to solve. Instead of buying off-the-shelf tools that almost fit, we built systems tailored to the exact requirements — and learned things in the process that directly benefit our client work.

Internal R&D keeps our team sharp on the technologies that matter: modern API design, distributed systems, AI orchestration, and security architecture. When a client engagement requires these capabilities, we are not learning on their dime — we have already built it, broken it, and rebuilt it.

Interested in what we're building? Let's talk.

Whether you need a platform built from scratch, an architecture designed for scale, or a team that has already solved problems like yours — we are ready.

Start the Conversation