The Work
Compilers, AI infrastructure, and the systems underneath.
Low-level rigor meets product pressure. From LLVM internals to multi-cloud ML pipelines — I build the parts that can't be shallow.
DiffSwarm
Multi-agent CLI that reviews pull requests using a structured propose, refute, and synthesize workflow
PythonFastAPIMCPMulti-AgentCLI
Tenstorrent TT-Metal (Open Source)
TTNN model bring-ups for modern vision architectures on Tenstorrent accelerators
TTNNTT-MetalAI AcceleratorsOpen Source
Keyda AI Tutoring Platform
Production tutoring workflows built on retrieval, guardrails, and multi-tenant AWS infrastructure
RAGAWSpgvectorNext.jsMulti-tenant
DecorateAI GenAI Platform
Real-time visual-processing inference prototypes with ARKit integration
GenAIComputer VisionARKitModel Optimization
RISC-V Performance Optimization
LLVM/Clang tuning for RISC-V RVV vector workloads at SiFive
LLVMClangRISC-VVectorizationCode Generation
Swift-C++ Interoperability
Led interoperability initiative across Swift/Clang toolchain at Apple
SwiftC++ClangCompilerABI
The Athlete
Outside work, same operating system.
0.0
Marathon Miles
0.0 km
Olympic Triathlon
0
Podium Finishes
0+
Meters Elevation

Race Telemetry — Live Data
Loading telemetry...
The Proof






