Insights & Writing
Curated technical articles on backend architecture, system design, modern fullstack development, and production engineering practices.
A production-grade breakdown of GLM-Legal: how to adapt a general GLM-style decoder into a citation-grounded legal assistant using continued pretraining, LoRA SFT, hybrid retrieval, and constrained decoding. Includes runnable Python, latency/cost trade-offs, and a deployment checklist.
A data-driven engineering analysis of PlayStation 5 pricing from 2020 to 2026. We build a Python model to quantify inflation, semiconductor supply chain shifts, and the economics of console hardware.
TrailSafe is an open-source, offline-first safety system for solo hikers that fuses on-device TabPFN tabular inference with temporal sequence models to detect distress when cell signal dies. This deep dive covers the architecture, a production-grade Python implementation, and the latency/accuracy/cost trade-offs of running AI at the edge.
A deep dive into LineageOS 24.0's architecture: how the fork rebases on Android 16 QPR, enforces Treble VNDK 35, migrates to kernel 6.1 LTS, and ships a hardened, reproducible build pipeline for 200+ devices.
I benchmarked 13 production LLMs on a controlled knowledge-conflict task: the model is told a new entity name in-context, yet 12 of 13 still emitted the stale pretrained name. This article dissects the mechanism, ships a runnable harness, and quantifies the latency/cost trade-offs of mitigation strategies.
A production-grade architecture for redacting PII entirely in the browser using WebAssembly, the Shape Detection API, and Web Workers. Covers the memory model, latency budgets, and the trade-offs of moving redaction from server to client.