Hi @LeoGao, I just submitted The Manhattan Kernel to your microgrant program. It is a strict, CPU-bound Rust runtime architecture that treats the LLM purely as an isolated hypothesis generator, driving execution in graph-space to maximize local LLM avoidance. We already have 25 green integration tests in our open-source repo. I would love for you to check out our architecture!
László Lipcsik
@lipcsiklaci
Systems engineer and independent AI safety researcher focused on deterministic runtime architectures and low-level optimization. Specialized in Rust and hardware-constrained environments. Currently developing bottom-up approaches to enforce behavioral boundaries on large language models through structured execution spaces and local verification loops, shifting the agent paradigm from token-space reasoning to verifiable graph-space architectures.
$0total balance
$0charity balance
$0cash balance
$0 in pending offers
About Me
I am a systems engineer working independently on deterministic AI runtime architectures. My background is in low-level software design and hardware-constrained development, moving away from high-level application wrappers toward strict execution sandboxes.
I focus on building production-grade scaffolding that treats large language models as untrusted, non-executable operators. Rather than trying to align neural networks through prompt optimization, my research focuses on building rigid external constraints in Rust—such as state-machine verification and graph-mutation loops—to ensure 100% predictable system behavior.
My work is strictly code-first and open-source, verified by integration tests and real-world compiler feedback rather than theoretical papers.