Constance Angelopoulos
Controls Engineer at Harbinger
Berkeley, California, United States
Summary
🤩
Rockstar🎓
Top SchoolConstance Angelopoulos is a Controls Engineer based in Berkeley focused on medium-duty EV truck systems at Harbinger Motors, bringing six years of combined industry and academic experience in automotive and aerospace control systems. She holds an MS in Control Theory from UC Berkeley and has hands-on expertise implementing MPC, LQR, PID, and state estimation in simulation and vehicle software, plus MATLAB/Simulink toolchains including automatic code generation. Her contributions span vehicle state estimation, cruise control design, and efficient CAN-log processing that sped up debugging cycles during her Harbinger internship. As an MPC researcher she compared nonlinear MPC and LQR controllers for aggressive trajectory tracking and built visualization and optimization tools for trajectory generation. She also contributes to ML tooling on GitHub, implementing conformal prediction workflows for image classification uncertainty and exploring distribution-shift robustness. Comfortable bridging theory and production, she blends rigorous academic training with pragmatic engineering that shortens the path from control design to on-vehicle deployment.
6 years of coding experience
3 years of employment as a software developer
Master of Science - MS, Mechanical Engineering, 4.0, Master of Science - MS, Mechanical Engineering, 4.0 at UC Berkeley College of EngineeringHigh School Diploma, High School Diploma at Polytechnic School
Stackoverflow
11reputation
1kreached
2answers
1question
Github Skills (13)
scikit-learn10
machine-learning10
python10
image-classification10
numpy10
scikit10
computer-vision9
css6
nvidia6
opencv6
android6
tensorflow6
gpu6
Programming languages (7)
Github contributions (5)
aangelopoulos/conformal-prediction
Dec 2021 - Jan 2023
Lightweight, useful implementation of conformal prediction on real data.
Role in this project:
ML Engineer
ML EngineerContributions:136 commits, 4 PRs, 21 pushes in 1 year 1 month
Contributions summary:Constance implemented several scripts and notebooks related to conformal prediction, a method for uncertainty quantification. They created the initial scripts for the Lightweight Ambiguous Classifier (LAC) and generated ImageNet scores and labels using ResNet152, showing a clear focus on image classification and related techniques. Furthermore, the user explored several methods like the RAPS algorithm to assess and improve the accuracy of the prediction sets. The commits indicate work on distribution shift and analysis of the quality and size of resulting prediction sets.
predictiondistribution-shiftinferencecomputer-visionuncertainty-quantification
aangelopoulos/conformal_classification
Sep 2020 - Jan 2023
Wrapper for a PyTorch classifier which allows it to output prediction sets. The sets are theoretically guaranteed to contain the true class with high probability (via conformal prediction).
Contributions:147 commits, 4 PRs, 36 pushes in 2 years 5 months
classifiernonparametric-statisticstorchvisionimage-classificationclassification
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.