Digital Instruments
& Case Studies.

Cut my photo-restoration platform's GPU costs by ~73% while beating commercial tools on quality. The naive multi-API approach (separate services for upscaling, face restoration, and colorization) was expensive and inconsistent, so I engineered a unified orchestration layer managing 14+ deep learning models (SUPIR, HAT, CodeFormer) with thread-safe VRAM allocation and LRU eviction across 49GB. Delivers 28.5dB PSNR at 99.95% uptime... outperforming Magnific AI and Topaz on blind tests.

A cybersecurity LLM project: a 442,000-example instruction dataset and the synthetic-data pipeline behind it (tool-calling traces, ReAct reasoning, OPSEC scenarios, automated quality filtering), plus a two-stage LoRA fine-tuning setup for Qwen2.5-Coder-32B. Stage-one domain adaptation was trained to completion on rented 8x A100 GPUs; the dataset, pipeline, and training configs are the shipped artifacts.

VoiceKeep (shipped as voicekeep.io) grew out of the Voice Cloner research prototype ... originally developed on an RTX 3080, now running on a dedicated GPU server with RTX PRO 6000 Blackwell ... into a production AI voice platform handling single-voice TTS, multi-speaker conversations, and full audiobook production from manuscript uploads. The platform runs Qwen3-TTS 1.7B with 12-second P50 latency, 41+ curated voices, and zero-shot cloning from short reference audio. The Audiobook Studio parses DOCX/PDF/TXT manuscripts into chapters with dialogue detection, assigns character voices, applies pronunciation dictionaries, and exports distribution-ready M4B with chapter markers. Multi-voice conversations support drag-and-drop line ordering, per-line effects (speed, volume, gap), stage directions, multiple takes, ambient audio, and a waveform timeline editor. Ran with Stripe subscription billing on a single dedicated GPU server; the hosted service has since been retired, but the platform is a full worked example of production TTS engineering.

FastAPI + Next.js 15 SaaS boilerplate with 200 tests at 98% coverage and a 22-item OWASP ASVS Level 1 security checklist. Built because shipping the same boring SaaS infrastructure (auth, billing, deploy configs, security hardening) shouldn't cost a weekend per project. Handles the unglamorous 80%... Clerk JWT, Stripe webhooks, async SQLAlchemy, Docker deploys... so you can build the actual product. Sold on Gumroad. Pro $149, Agency $499.

Cut vector conversion time from 45 minutes to 8 seconds per asset... a 337x speedup. Design teams were hemorrhaging billable hours manually tracing logos and icons in Illustrator. Built a GPU-accelerated pipeline combining neural upscaling with dual vectorization engines (Potrace + VTracer), plus an SVGO optimization stage that reduces file sizes by 40-60%. Now processing 2,000+ conversions monthly with zero manual intervention.

Recovered 2+ hours daily lost to context-switching between terminal, database clients, and config files. Claude Code power users were drowning in fragmented tooling... no unified view of sessions, memory state, or MCP server health. Architected a native Electron control center with 25 tRPC endpoints managing PostgreSQL, Memgraph, and Qdrant memory systems. 80% test coverage, zero production incidents since launch.

Built a generative-engine optimisation auditor, then ran it against alexmayhew.dev and published a summary of that run without rerunning or rescoring it. Site score 46, cited in 3 of 18 observations across 2 of 4 surfaces, 44 of 44 technical checks passing. Includes the defects the tool itself got wrong: a coverage false negative and a misattributed citation basis.