> long-form thoughts on software, ai, and building products.
A practical guide to fine-tuning Qwen3.5-4B on X and LinkedIn writing data using Unsloth QLoRA, then evaluating baseline vs fine-tuned outputs with llama.cpp.
A deep dive into building Clawbench, an open-source benchmarking tool for evaluating LLM performance across real-world tasks. Learn how to systematically test and compare AI model outputs.
How I built Demand Signal, an AI-powered platform for tracking and interpreting market demand signals across data sources. A technical walkthrough of the architecture and lessons learned.
A curated list of the best open source AI tools for developers in 2026. From LLM evaluation with Clawbench to market intelligence with Demand Signal — the tools I use and build every day.
> external writing
2026 Here I Come↗ Substack