Compute only
what matters.
I build working proof-of-concepts for difficult computational problems — then test them against measurable results.
Completed PoCs.
Four different classes of systems problems: selective computation, incremental state, multi-engine orchestration and local-state simulation.
TESERAKT
Selective Computation Engine
Find the smallest dependency set required for a target result and avoid executing unrelated work.
TIME
Incremental State Engine
Preserve work that is still correct. Recompute only the part actually affected by change.
VM
Virtual Motherboard
Route work across independent engines while keeping their validated cores isolated.
WORLD 1.0
Local-State Simulation Foundation
A local event should activate local consequences — not force a global world update.
Small engines. Clear jobs.
Supporting modules remain separate from the featured systems and are presented as capabilities, not inflated into products.
TUNNEL
Verified result retrieval and controlled execution path.
DELTA
Represent the change, not the whole state.
VECTOR
Control where work propagates and reject unrelated workload.
Evidence before claims.
A prototype is not successful because it runs. The expected behavior must survive the relevant execution test.
Give me the difficult part.
If your team has a computational bottleneck, an unusual systems problem or an idea that needs a working PoC, define the problem and what would count as a useful result.