Engineering studies.

Live Projects

MLOps & AI Infrastructure

Built Bayesian neural network inference pipelines to quantify prediction uncertainty in production ML workflows. Scaled high-dimensional SVM models using kernel approximation to reduce compute cost. Developed a modular optimization framework comparing adaptive and second-order gradient methods.

  • PyTorch
  • MLOps
  • Kubernetes
  • AWS

Fault-Tolerant Distributed Systems

Designed and evaluated Paxos, Raft, and Byzantine consensus protocols for leader election and fault tolerance. Built an event-driven transaction system using vector clocks and CRDTs for consistency under failure.

  • Go
  • Distributed Systems
  • Consensus
  • CRDT

FPGA-Accelerated AI Systems

Implemented an FPGA-based inference accelerator using low-precision quantization and sparsity. Designed a high-speed SPI controller with correct clock-domain crossing and synchronization.

  • FPGA
  • Verilog
  • AI Acceleration
  • Hardware
View All on GitHub