How I Made CosyVoice3 2.2× Faster on Apple Silicon
How I bounded PyTorch MPS memory growth, cut CosyVoice3 latency below real time, and moved only the dispatch-bound LLM stage to MLX.
Category archive
Software Engineering
Field notes collected under Software Engineering.
25
entries
→
One Mattermost Bridge, Two Hot-Swappable Agents
How I separated Codex and Pi from a growing Mattermost bridge, removed duplicated agent plumbing, and made each runtime replaceable without stopping routines.
How to Benchmark AuthOS in a 1 vCPU, 1 GiB VM
Build a constrained KVM guest, run the AuthOS SQLite workload from the host, collect comparable results, and select a load for a longer test.
From Noisy Kenyan ID OCR to JSON with MLX and Candle
Fine-tune a 50M-parameter model on synthetic Kenyan ID OCR, measure field-level accuracy, and package the fused F16 weights for Candle inference in Rust.
From PP-OCRv6 Text to Kenyan ID JSON
A measured walkthrough of the modes, schema, synthetic fixture, and small understanding model I put around PP-OCRv6 to extract Kenyan ID fields locally.
Schedule Codex Routines as Finite Runs
How my Mattermost bridge keeps recurring schedules outside Codex, gives each run explicit state, and records the result without leaving an agent goal alive.
How I Run Codex from Mattermost on My Phone
The architecture, bridge, and security boundary behind my self-hosted Mattermost interface to Codex on a local development machine.
Why I Built a Desktop UI for Perplexity's Bumblebee Scanner
I built a Flutter desktop UI for Perplexity's Bumblebee scanner to make scan scope, package inventory, exact exposure matches, diagnostics, and history easier to inspect.
Building a Review-Focused Flutter Client for Jules
How I built a Flutter client that keeps Jules sessions, approvals, and code diffs in one workspace, with account-scoped caching and visible sync state.
Building Dream TTS Around Durable Local Speech Jobs
How Dream TTS separates its Flutter interface from Rust synthesis work so document imports, resumable jobs, model files, and private audio stay manageable.
Testing Local TTS: Pronunciation, Latency, and Release Gates
A practical quality loop for local text-to-speech: representative fixtures, contextual normalization, useful latency metrics, defect labels, and release checks.
Designing a Stable API for Local Text-to-Speech
A practical design for keeping engine selection, sentence queues, playback, voice packages, and recovery behind one local speech service.
