Gihyun Shim

About

UPenn robotics researcher who architected a real-time autonomous driving decision-making engine, integrating LSTM trajectory prediction with MPC in CARLA and adding conformal prediction to deliver 95% statistical safety guarantees under strict latency constraints. Also built and debugged an autonomous quadrotor stack with ESKF-based 6-DoF tracking and optimized A*/Dijkstra planning to eliminate latency-induced instability, with experience bridging heterogeneous simulation/control systems.

Experience

Education

Key Strengths

• Architected real-time autonomous decision-making engine for dynamic driving environments
• Built end-to-end simulation-to-control pipeline integrating LSTM trajectory prediction with MPC in CARLA
• Achieved 95% statistical safety guarantees using conformal prediction + MPC
• Optimized real-time control loop by reducing conformal prediction overhead (sliding window + sampling strategy)
• Rapidly adapted ROS-style architecture for quadrotor navigation (high-frequency messaging, TF2 transforms)
• Debugged and stabilized obstacle avoidance by profiling and optimizing A* (local-window search) to eliminate latency spikes
• Improved flight stability via ESKF covariance re-tuning for smooth, drift-free 6-DoF tracking
• Designed custom communication layer bridging heterogeneous platforms (Unreal/CARLA to Python control engine)
• Created stochastic simulation scenarios to stress-test safety and reliability prior to hardware integration