Background
Mike started as a hardware technician working on IBM minicomputers in the 1970s, moved into professional operating-system development in the 1980s, worked on research technology-transfer middleware in the 1990s, and then spent the 2000s and onward at many Silicon Valley startups.
Current Work
Current projects use historically grounded systems such as sw-comp-history and sw-embed as demanding testbeds for developing AI coding-agent tooling and for fine-tuning local LLMs.
The practical thread is straightforward: build real emulators, assemblers, compilers, demos, and documentation; then use those projects to find out where agent workflows, local models, and supporting tools still need to improve.
Career Arc
- 1970sHardware technician working on IBM Series/1, 1130/1800, System/3, System/32, System/34, 5100/5110/5120, and System/7 machines.
- 1980sOperating-system developer on IBM MVS/370, MVS/XA, and MVS/ESA components including Dynamic Reconfiguration and Standalone Dump.
- 1990sResearch technology-transfer middleware programmer working around Automatic Restart Manager, Distributed Syncpoint Manager, DSOM, ILM, IMS, Java Record I/O, MVS/BatchPipes, Sysplex, and Workload Manager.
- 2000s onwardEngineering roles at many Silicon Valley startups, including 4GL, platform engineering, release engineering, management, mentoring, React development, and AWS.
- NowRust programming, machine learning, AI coding-agent tools, local LLM fine-tuning, and historically grounded software projects.
Hobbies
Electronics, PCs, Linux, embedded systems, homelab infrastructure, machine learning, and AI.