Projects

Featured technical work

Work spanning compilers, AI infrastructure, and systems engineering— from Swift/C++ interoperability to RISC-V vectorization and resilient AI platforms.

Tenstorrent TT-Metal (Open Source)

TTNN model bring-ups for modern vision architectures on Tenstorrent accelerators

TTNNTT-MetalAI AcceleratorsOpen Source
  • MaskFormer Swin-B hybrid decoder + perf artifacts (PR #32335)
  • DPT-Large depth estimation bring-up (PR #33123)
  • YOLOS-small object detection bring-up (PR #32500)

Keyda AI Tutoring Platform

Production tutoring workflows built on retrieval, guardrails, and multi-tenant AWS infrastructure

RAGAWSpgvectorNext.jsMulti-tenant
  • Teacher-orchestrated learning workflows with retrieval + context
  • Secure multi-tenant architecture with RBAC and audit logging
  • Evaluation harness, structured outputs, and deployment automation

DecorateAI GenAI Platform

Real-time visual-processing inference prototypes with ARKit integration

GenAIComputer VisionARKitModel Optimization
  • Low-latency inference pipeline with caching and throughput tuning
  • Evaluation framework for prompt/model changes
  • ARKit-integrated prototypes for interior-design workflows

RISC-V Performance Optimization

LLVM/Clang tuning for RISC-V RVV vector workloads at SiFive

LLVMClangRISC-VVectorizationCode Generation
  • Tuned vectorization cost models
  • Architecture-specific optimizations
  • Built CI/CD for cross-compilation

Swift-C++ Interoperability

Led interoperability initiative across Swift/Clang toolchain at Apple

SwiftC++ClangCompilerABI
  • AST bridging and name mangling
  • Coordinated Apple Silicon migration
  • Open-source contributions

Let’s build something ambitious

I partner with teams shipping compilers, AI systems, and mission-critical infrastructure.

Start a conversation