ML Engineer focused on LLM post training, and inference infrastructure.
Coding implementations are built upon numerous implicit strong assumptions. This explains why iterative simulation and code tuning can never replace rigorous theoretical reasoning. Writing clean code is rewarding, yet never overlook the fundamental connections beneath algorithmic updates.
- Machine Learning & Game AI
- Building LLM-powered agents and multi-agent systems; researching frontier model mechanisms (MLA, MoE, DSA, DSpark)
- Languages: C++ · Python · Java · JavaScript
fooSynaptic — foo from the programmer's foo / bar: machine learning is, at heart, learning to solve for the unknown; Synaptic for intelligence wired like synapses.



