Building for efficiency

Product Manager & AI Integration Specialist

A second brain approach combining Notion databases and memory.json knowledge graph for effective project management across multiple domains. I design hybrid systems that blend AI capabilities with traditional workflows.

My Skills

Product Strategy

Roadmapping, A/B Testing, Project Management, and Strategic Planning for complex product ecosystems.

Product Design

User Stories, User Flows, Wireframes, Persona Development, and UI/UX Implementation.

Market & User Research

Market Analysis, Experiment Design, Usability Testing, and Competitive Analysis.

AI Integration

LLM Implementation, Knowledge Graph Design, Hybrid AI/Human Systems, and Token Optimization.

Featured Projects

Hybrid PM Implementation

A comprehensive project management system that combines Notion databases with memory.json knowledge graph for enhanced project tracking and context awareness.

Notion Integration Knowledge Graph AI
View Project

Memory Visualization System

Interactive 3D visualization of a knowledge graph that enables intuitive exploration of entity relationships and system components.

D3.js Three.js Visualization
View Project

Fallback OpenAI Client

Resilient AI system that automatically switches between service providers during outages, with 36% token optimization and confidence scoring framework.

Upsonic LLM Resilience
View Project

System Architecture

Notion Database Projects Database Tasks Database Sprints Database WOW Review System memory.json Knowledge Graph Project Entities Task Entities Observations Relations Synchronization Layer Entity Mapping | Conflict Resolution | Change Detection | Data Transformation Access Layer Claude AI | CurserAI | Query Builder | Response Templates

Notion Integration

The system utilizes Notion databases for detailed project tracking, task management, and sprint organization. Key components include Projects, Tasks, and Sprints databases with bidirectional relationships.

Knowledge Graph

memory.json provides a flexible knowledge graph architecture for storing entity relationships, observations, and quick-access context. This component enables advanced query capabilities and context-aware filtering.

WOW Sprint Review

The Week of Work (WOW) Sprint Review system enables consistent assessment of sprint progress across multiple projects. It follows the rule of always reviewing the latest sprint before resuming work on any project.

Sprints Management

The system implements a cross-project sprint approach where tasks from different projects can be grouped into common sprints. This enables efficient management of multiple projects simultaneously.

Sprint 1: Project Setup

March 25-27, 2025
  • Project creation in Notion
  • memory.json server setup
  • Entity mapping implementation
  • CRUD operations testing

Sprint 2: Notion Enhancement

March 28-31, 2025
  • Kanban view implementation
  • Calendar view configuration
  • Enhanced table view
  • Custom filtered views

Sprint 2: Memory Structure

March 28-31, 2025
  • Entity schema design
  • Relationship mapping
  • Observation standardization
  • Naming conventions

Sprint 3: Visual Improvements

April 1-6, 2025
  • Template management
  • Visual element standards
  • Chat context solution
  • MCP research

Sprint 4: MCP Implementation

April 7-14, 2025
  • Model Context Protocol
  • External integrations
  • Advanced queries
  • System productization

Sprint Database Implementation

Cross-Project Capabilities

The Sprint Database allows tasks from different projects to be organized into common sprints, enabling efficient management of parallel work streams.

Bidirectional Relationships

Established relationships: Project:Task (One:Many), Sprint:Task (One:Many), Project:Sprint (Many:Many) for complete traceability.

WOW Review Process

Structured approach to reviewing sprints with standardized templates and assessment criteria. Always reviews latest sprint before resuming work.

Upsonic Fallback System

Our resilient AI-powered fallback system automatically switches between service providers when one becomes unavailable, ensuring continuous operation of the PM system.

Upsonic Fallback System Architecture 36% Token Optimization with Automatic Service Switching Client Request Upsonic Router Service Monitoring Token Management Anthropic API (Primary Service) OpenAI API (Fallback Service) Response Processor Format Standardization Optimized Response Token Optimization Layer Context Management | Content Chunking | Priority Routing | 36% Token Reduction Confidence Score 0.8

36% Token Optimization

By implementing token-aware architecture at the system level, we achieved a 36% reduction in token usage while maintaining response quality, resulting in faster responses and lower costs.

Confidence Scoring System

A 0.0-1.0 scale confidence scoring framework evaluates reliability across tools, processes, and tasks, providing crucial metrics for prioritizing fallback triggers and system improvements.

Automatic Service Switching

The system continuously monitors service availability and automatically routes requests to alternative providers when needed, ensuring zero downtime operation even during cloud provider outages.

Implementation Details

Sequential Thinking Approach

Complex implementation challenges were broken down using a structured sequential thinking methodology, enabling systematic problem-solving even for difficult integration issues.

Model Mapping

Intelligent mapping between different service providers ensures consistent response formatting regardless of which backend is handling the request.

Rapid Debugging

A streamlined diagnostic pipeline quickly identifies and resolves potential issues like missing API credentials before they can cause cascading failures in the system.

Project Progress

Overall Progress 42%

Sprint 1: Project Setup 100%

Sprint 2: Notion Enhancement 65%

Sprint 2: Memory Structure 70%

Sprint 3: Visual Improvements 15%

Recent Achievements

Upsonic Fallback System

Successfully implemented a resilient fallback system using the Upsonic framework, enabling automatic service switching and achieving 36% token usage reduction through optimization.

Sequential Thinking Tools

Integrated MCP sequential thinking tools for enhanced problem-solving capabilities, providing structured approaches to complex implementation challenges.

Confidence Scoring Framework

Developed a 0.0-1.0 scale confidence scoring system to evaluate reliability across tools, processes, and tasks, providing crucial metrics for system improvements.

Next Steps

Layer 2 Memory with Notion

Implementing an enhanced Layer 2 memory system using Notion as a structured database layer on top of memory.json, creating a dual-layer architecture for improved memory management.

Sprint Process Enhancements

Improving the sprint review process with automated metrics collection and performance analysis using our confidence scoring framework to identify optimization opportunities.

Upsonic Testing Framework

Developing a comprehensive testing framework for the Upsonic fallback system to ensure reliability through simulated service disruptions and edge case handling.

Layer 2 Memory Architecture

The Layer 2 Memory system enhances our knowledge graph with structured database capabilities, creating a dual-layer architecture for optimized memory management.

Layer 1: memory.json Knowledge Graph Entities Relations Observations Priority Classification System Layer 2: Notion Workspace Projects Database Tasks Database Sprints Database WOW Sprint Review System Bidirectional Sync Layer

Layer 1: Knowledge Graph

The memory.json system forms our foundational layer, providing flexible entity relationships, observations storage, and a comprehensive priority classification system for knowledge evaluation.

Layer 2: Structured Database

Notion workspace provides structured database capabilities with Projects, Tasks, and Sprints databases, along with our WOW Sprint Review system for consistent assessment and tracking.

Bidirectional Sync

The sync layer ensures changes in either system are properly reflected across both layers, maintaining consistency while leveraging the strengths of each platform.

Key Benefits

Enhanced Context Persistence

By distributing memory across two complementary systems, we achieve better context persistence across conversations and sessions, addressing the chat context limit challenge.

Optimized Query Performance

The dual-layer approach allows for optimized queries - using memory.json for relationship-focused lookups and Notion for structured, property-based filtering and sorting.

Fallback Reliability

With critical information stored in both systems, the architecture provides natural fallback capabilities if either system experiences temporary availability issues.

Limbic System

Manages relationships and connections
Relationship Mapping
Connection active
Frontal Lobe (Memory System)
Parietal Lobe (Processing)
Temporal Lobe (Notion DB)
Occipital Lobe (Visualization)
Cerebellum (Task Execution)
Limbic System (Relationships)
Click and drag to rotate brain • Zoom with mouse wheel • Click on brain regions to explore components
System Active
Initializing Second Brain...