Kieran Lachmansingh

About Me

Hello! I'm a passionate robotics and high-performance computing graduate exploring the frontier between intelligent algorithms, real-time systems, and scalable design.

I'm currently finishing my Master's in Applied Science at Queen's University. My journey includes applying data analysis tools like Excel, programming design and testing using ROS and ROS2 for robotics, and experience working with data center-sized systems.

With a background in electrical engineering and hands-on experience in Python, C, and C++, I aim to combine technical depth and creative problem-solving to develop high-impact solutions in robotics and high-performance computing.

Affiliations

MItHRIL

Ingenuity Labs Research Institute

CAESAR

Research & Publications

Selected research projects at the intersection of robotics, real-time systems, and high-performance computing.

SHARP: Supercomputing for High-speed Avoidance and Reactive Planning

ICRA 2026

Kieran Lachmansingh, José R. González-Estrada, Jacob Chisholm, Ryan E. Grant & Matthew K. X. J. Pan

SHARP investigates whether high-performance computing can be used for reactive robot control without violating real-time constraints. Motion planning is offloaded to HPC clusters using a hash-distributed, multi-goal A* planner. The system enables millisecond-scale avoidance behaviors and reframes HPC offloading as a viable pathway toward dependable, reactive robots in dynamic environments.

Project page Paper Code

A computer connected to several servers.

Scaling Parallel Graph Computing for Real-time Robotics

ICS 2026 (in review)

Kieran Lachmansingh, José R. González-Estrada, Jacob Chisholm, Matthew K. X. J. Pan & Ryan E. Grant

This paper presents a systematic scaling analysis of a repre- sentative robotics planning algorithm, motivated by the emerging need for robots to leverage data center-scale compute resources. To study this interaction, we implement a hash-distributed multi- goal graph search that computes collision-avoidance trajectories. We evaluate its performance under stress, examining both strong and weak scaling across a range of graph sizes and process counts.

Project page Code