Simulation & Applied AI

Latency-aware MR training systems.

Instrumented simulators and real-time pipelines—tracking, rendering, metrics, evaluation.

MR surgical simulators blending physical instruments + virtual environments

Real-time CV inference with latency-aware UX decisions

Patents/publications; delivered on PI-led funded research as engineering contributor

Demo Highlights

Real-time overlays showing tracking confidence, error, and latency to keep operators informed.

Budgeting render, IO, and inference to stay within comfort thresholds for trainees.

Session recording with scoring hooks for studies and skill assessment.

Case Study 1

Role: Role: simulation engineer (pipeline design, instrumentation, performance).

Context: C++/Unity simulator blending physical instruments with virtual render and tracking.

Problem: Problem: need realistic training with measurable performance and low latency.

Constraints: Constraints: latency budget across render/IO/inference, tracking accuracy, calibration reliability.

"Impact: demo-ready for studies, clear metrics for evaluator review, path to publication."

Instrumented render + tracking pipeline with budgets per stage and alerts on spikes.
Calibration tooling to align physical instruments with virtual overlays (VTK/OpenCV).
Metrics capture (task time, errors, smoothness) stored for evaluation.
Profiling loops to keep frame times inside comfort thresholds.

Artifacts: demo reel, calibration scripts, metrics schema, performance profiles.

Latency budgets: track where time is spent (render, IO, inference, tracking).
Instrumentation-first: measure before optimizing; surface spikes to the operator.
Designing for uncertainty: sensor noise, inference confidence, and fallbacks.
Evaluation: metrics that map to training outcomes and study protocols.

PI-led funded programs (engineering contributor)
ARG Cycle 2 (2025–2028) — USD 739K
ARG Cycle 1 (2024–2027) — USD 739K
QNRF NPRP 11 (2019–2023) — USD 773K
QNRF NPRP 05 (2016–2019) — USD 1.01M