Agent Skills spec + Mastra integration
· 9 min read
Senior Software Engineer building AI-powered products with TypeScript, Rust, and LLMs. Writes about AI agents, eval-driven development, and edge computing.
View all authorsAI-Powered Skill Extraction with Cloudflare Embeddings and a Vector Taxonomy
· 4 min read
This bulk processor extracts structured skill tags for job postings using an AI pipeline that combines:
AI Observability for LLM Evals with Langfuse
· 10 min read
This article documents an evaluation harness for a Remote EU job classifier—but the real focus is AI observability: how to design traces, spans, metadata, scoring, and run-level grouping so you can debug, compare, and govern LLM behavior over time.
Schema-First RAG with Eval-Gated Grounding and Claim-Card Provenance
· 7 min read
This article documents a production-grade architecture for generating research-grounded therapeutic content. The system prioritizes verifiable artifacts (papers → structured extracts → scored outputs → claim cards) over unstructured text.
You can treat this as a “trust pipeline”: retrieve → normalize → extract → score → repair → persist → generate.
Evals for Workflow-First Production LLMs: Contracts, Rubrics, Sampling, and Observability
· 12 min read
Agentic Job Pre-Screening with LangGraph + DeepSeek: Auto-Reject Fake “Remote” Roles
· 7 min read
Introduction
Building Long-Running TTS Pipelines with LangGraph: Orchestrating Long-Form Audio Generation
· 17 min read
Introduction
DeFiTuna: On-Chain Limit Orders on Solana
· 14 min read
TL;DR
Deploying Real-Time Solana Data Streams on Cloudflare Containers with LaserStream
· 12 min read
DefiTuna AMM Smart Contract — Anchor Implementation Guide
· 11 min read