How Local LLM Throughput Scales with Multiple Agents
Adding agents does not simply divide a local LLM’s solo token rate between them. Total output can rise even while each agent slows down.
Read articleI test backend systemsunder load and failure.
SDET at MariaDB, working on distributed systems, performance, and test infrastructure. Outside work, I experiment with local LLMs and build agent tools.
- Now
- SDET at MariaDB — distributed systems, performance, test infrastructure.
- Shipping
- iPhone apps I design, build, support and publish myself. See them
Latest writing
All writingAt work
Databases, distributed systems, and performance.
Backend and databases
Correctness and integration testing against real dependencies — the failures that only appear once the data is production-shaped.
Performance and distributed systems
Load, contention, and race conditions: behaviour no component shows while it is tested on its own.
AWS and test infrastructure
Environments, CI/CD, and observability that make a rare failure repeatable enough to be worth fixing.
Selected projects
Research and engineering tools.
Experiments with explicit limits, tools for practical work, and infrastructure for testing it.
01compact_llm_languageCompleted research · local LLMsControlled experiments in hidden-state recurrence for Qwen3.5-2B. The tested recipes produced no replicated reasoning gain; the methods, controls, and negative results are documented.02requirements-minerAgent tooling · requirementsRequirements tooling that runs isolated expert reviews, weighs evidence, and preserves disagreement. One shared skill across multiple agent harnesses.03serbian-dubPython · end-to-end media pipelineA Russian-to-Serbian video dubbing pipeline: transcription, speaker-aware translation, voice synthesis, timing, mixing, and automated QA.
More projects on GitHub iPhone apps
Apps I design, build, support and publish myself, each one doing a single job for a specific person.
View all apps