Writing.
I write about the software development process and the technical challenges I face daily as a software engineer.My writing draws from hands-on experience building mastra.ai, an all-in-one Typescript-based framework for building AI-powered applications and agents.
Workflow Orchestration: Event-Driven Architecture & HITL Patterns
How we built sophisticated AI workflow orchestration with streaming, suspension/resumption, error handling, and human intervention points.
AI Evaluation Pipelines: Creating Custom Scorers for LLM Quality Assessment
How we built a comprehensive evaluation system that balances deterministic algorithms with AI-powered judgment using a four-step pipeline approach.
Tool-Calling Accuracy in AI Agents: Building Reliable Multi-Tool Systems
How we built a comprehensive tool accuracy evaluation system that measures and improves AI agents ability to select and execute the right tools in complex multi-tool environments.
Building Context-Aware AI: The Science Behind RAG Quality Metrics
How we built sophisticated metrics to measure context quality, ranking algorithms, and the nuances of information retrieval for AI systems.
From Code to Production: Building Ready-to-Deploy AI Application Templates
How we created 12+ production-ready templates that democratize AI application development with real-world complexity and deployment patterns.
Multi-Modal AI Applications: From PDF Processing to Audio Generation
How we built production multi-modal AI pipelines that seamlessly transform content across different modalities with robust error handling and resource management.