Skip to content

Projects

Work across AI systems, control and robotics, optimization and industrial data science. Each write-up covers the problem, the approach taken, and what the result does and does not show.

Top-down visualization of estimated camera poses arranged around the calibration board, showing where each shot was taken from.

Camera Calibration and NeRFStudio Pipeline

Python pipeline that calibrates a camera from an ArUco board, estimates poses, and converts them into the transforms NeRFStudio expects for 3D reconstruction.

  • Python
  • OpenCV
  • ArUco
Results section of the Embedding Lab report with recall@10 and all overlap tiers selected. A table of nine methods gives recall@10 with a bracketed interval, a bar with a line for the interval, and the difference to BM25 on full text. BM25 on full text has the longest bar, 0.69 [0.62, 0.76], and is the reference. Own-64 on full text is 0.44 [0.39, 0.49], 0.24 below it.

Embedding Lab

A 64-dimension embedding model trained from scratch in NumPy on NumPy and SciPy docstrings, compared with TF-IDF, BM25, LSA and static-vector baselines on the same text and queries; the trained model loses to BM25.

  • Python
  • NumPy
  • SciPy
Pipeline diagram running from a camera or FER2013 image through face detection and cropping, 48x48 grayscale normalization, HOG feature extraction and a scaler plus classifier, to an emotion label with a confidence value.

Facial Expression Recognition with Classical ML

Computer-vision pipeline that detects a face, extracts HOG features and classifies it into seven expressions, built as a transparent classical baseline rather than a deep-learning system.

  • Python
  • OpenCV
  • scikit-learn
Results chart for leave-one-participant-out validation, showing balanced accuracy per held-out participant and the spread between them.

Smartphone Activity Recognition from IMU Signals

Classifies human activity from phone accelerometer and gyroscope signals, with participant-level validation that shows how far the model actually generalizes to people it has never seen.

  • Python
  • NumPy
  • pandas
Convergence curve of best fitness against generation for the baseline genetic algorithm run.

Petri-Net Discovery with Genetic Algorithms

Evolves Petri-net structures from event-log traces with a genetic algorithm, comparing crossover, mutation and selection operators against a combinatorial search space.

  • Python
  • DEAP
  • pm4py
The dashboard shell styled as a retro Windows desktop, with stacked analysis windows open across the workspace.

Process Mining KPI Dashboard

Local process-mining dashboard for IT service ticket event logs, covering SLA monitoring, bottleneck detection, variant analysis and executive reporting.

  • TypeScript
  • React
  • Process Mining
Comparison plot of pitch and yaw tracking responses for the controllers in the benchmark, shown against the reference trajectory.

AERO Predictive Control Benchmark

MATLAB benchmark putting linear MPC, iterative nonlinear MPC, subspace predictive control and Koopman-lifted SPC on the same pitch and yaw tracking problem under one evaluation workflow.

  • MATLAB
  • YALMIP
  • MPT3
Simulated swarm trajectories for leader-follower flocking, with follower paths tracking the leader through the workspace.

Multi-Drone LTL Formation Control

MATLAB simulations that pair continuous LQR tracking with temporal-logic specifications, covering region-based planning, formation switching and leader-follower flocking for a small drone fleet.

  • MATLAB
  • LQR
  • Temporal Logic