Parth
Kheni

Machine Learning · Robotics · Computer Engineering

Currently

Beth Israel Deaconess Medical Center

Machine Learning Intern

Building GPU-enabled surgical robotics simulation workflows for the da Vinci Research Kit using ROS and Gazebo/AMBF. Improving ML autonomy training reliability by diagnosing preprocessing mismatches and curating demonstration data from JIGSAWS trajectories.

BU Center for Space Physics

ML Researcher

Developing LSTM models for space weather forecasting. Processing satellite telemetry data for predictive analytics of energetic electron precipitation events.

About

I'm a Computer Engineering student at Boston University with a concentration in Machine Learning. I take projects from data all the way to deployment, including training, evaluation, and integration.

My work spans surgical robotics simulation, space weather prediction, and embedded systems. I care about building systems that are easy to reproduce, measure, and improve over time.

ResumeGitHub →

Research Showcase

Aurora Flux Forecasting

LSTM-based machine learning models for predicting energetic electron precipitation events using POES satellite telemetry and OMNI solar wind data. This research aims to improve space weather forecasting accuracy for radiation belt dynamics.

View Project Proposal
E4_0 Channel (Smaller Dataset) - validation results

E4_0 Channel (Smaller Dataset) - Validation Set

Model predictions vs. actual energetic electron flux measurements from POES satellite data

View Full Size

7M+ Data Points

Processing 2-second satellite samples per month for time-series forecasting

85%+ Accuracy

LSTM model performance for energetic electron precipitation event forecasting

Multi-Input Architecture

Combining POES flux measurements with 4-hour windows of solar wind drivers

Technical Skills

Languages

PythonC/C++JavaJavaScript/TypeScriptMATLABVerilogSQL

ML/AI

TensorFlowPyTorchScikit-learnOpenCVPandasNumPy

Robotics

ROSGazeboAMBFArduino

Web/Cloud

ReactNode.jsAWSDockerGit

Interactive Demo

Neural Network Playground

Draw your own patterns and watch a neural network learn to classify them in real-time. A hands-on demonstration of machine learning fundamentals.

Try it out

Code Examples

Real Implementation Samples

Actual code from production projects. Click snippets to see different implementations.

Neural Network Training Loop
python
LSTM-based space weather prediction with custom loss function
Education
Boston University
BS Computer Engineering · Machine Learning
Expected Dec 2026
Rutgers University
BS Computer Engineering
Sept 2023 – May 2024
Get in Touch
I'm always open to discussing new opportunities, collaborations, or just connecting with fellow engineers and researchers.
pkheni@bu.edu
LinkedInGitHub
Get in Touch
Interested in collaborating or have a question? Feel free to reach out.
Or reach out directly via pkheni@bu.edu