Research Notes & Essays
Deep dives, engineering strategy, lessons from the field, and thoughts on building systems.
Quadrotor Surveillance: Local Peak Seeking and Smooth Coverage Control
An early research note on quadrotor surveillance — why plan-then-track and MPC-style solutions no longer felt like the right framing once the utility function was unknown, and why the real problem became turning local gradient information into smooth, trackable coverage behavior.
Dynamics of Optimization: From Faster Iterations to Faster Settling
An early research note on continuous-time optimization — why accelerated gradient methods no longer felt complete as discrete updates for autonomous systems, and why the real problem became designing the dynamics themselves for fast settling and possible circuit realization.
Lumbar Spine Classification: From One Global Model to Split Specialists
Reflections on why this lumbar-spine challenge worked better once the task stopped being treated as one unified classifier and started being split into more coherent target groups.
Designing the Markov-Matrix: Why Adapting a Random Walk Fell Short
An early research note on probabilistic swarm guidance — why adapting a random walk with Metropolis–Hastings no longer felt like the right answer, and why the real problem became designing faster-mixing Markov chains under graph and steady-state constraints.