Hi, I'm Essey Araya

Tools: Python, PyTorch, GPyTorch, SciPy

A Bayesian physics-informed ML toolbox for learning Port-Hamiltonian systems with uncertainty quantification, integrated into NeuroMANCER, PNNL's open-source scientific machine learning library.

This project includes:

• Gaussian Process kernel and mean function design for PHS structure
• Uncertainty-quantified parameter recovery across multiple physical systems
• Validation notebooks: inverted pendulum, DC motor, N-mass chain, driven Duffing oscillator
• 250+ unit tests, Sphinx/RST documentation pipeline
• Paper submission (IFAC format)

Tools: Next.js 16, TypeScript, Python, PostgreSQL, Vercel AI SDK

As Team Lead, I helped build Lumen, a tool for generating personalized degree plans for Berea College students.

My contributions include:

• A greedy prerequisite-graph solver in TypeScript that generated 4-year degree plans
• A course-ranking pipeline that generated 3 personalized degree pathways per student from career goals

Tools: Python, FastAPI, Gemini Vision API, Pydantic v2, Next.js 15

Built during a hackathon, SafeSource is an AI platform that quantifies PFAS exposure risk using EPA water datasets.

This project includes:

• A Python/FastAPI platform built in 24 hours to quantify PFAS exposure risk
• A Risk Exposure Index using geospatial, behavioral, and retail inputs to compute exposure scores
• A pipeline using Google Gemini Vision + Pydantic to estimate PFAS levels from grocery receipt images

Tools: Autodesk Inventor, Altair Inspire FEA

During the Fall 2025 semester, I led a team of 3 in redesigning a solar car chassis to reduce weight, and improve factor of safety and manufacturability.

I evaluated 18 load cases using FEA to identify safe material removal, then refined the geometry by communicating with the manufacturing team to ensure reliable manufacturing.

The final design achieved a 32% reduction in mass while fulfilling all safety requirements and improving manufacturability.

Tools: Composite Materials, Mold Making, Universal Testing Machine, Excel, ANSYS ACP & FEA, 3D Printing, Stress-Strain Curves

This project was completed during my summer 2025 internship at Berea College as part of the Solar Car team. I took ownership of designing and building a composite battery box and developed a scalable manufacturing approach for the solar car aeroshell.

I conducted coupon tests to evaluate strength to weight ratios of different composite layups and assess manufacturing quality, designed and fabricated female molds using 3D printed plugs. Design decisions balanced strength, weight, and manufacturability.

I achieved a lightweight composite battery box at 50% of the estimated budget and produced small scale aeroshell canopy.

Tools: OpAmps(LF353), Oscilloscope, Amplifier Circuit

An electronics project focused on measuring surface electromyography signals generated by muscle activity.

Utilized LF353 op amps to achieve sufficient gain while keeping noise low. I tested the circuit under different loading conditions to ensure the signal was stable and usable for analysis.

I chose electromyography for this lab project because of its impactful applications in prosthetics, medical diagnostics, and sport science.

Tools: PASCO Capstone, 3D Printing, Trigonometry, Mechanics

A video showcasing why a double cone appears to roll uphill. I show how the center of mass actually moves downward visually and mathematically, revealing the geometry behind the classic “uphill cone” paradox.

This Project includes:

• Video Motion Analysis using PASCO Capstone
• Mathematical proof of the change in position of the center of mass with respect to time
• Lagrangian of the System

Tools: Autodesk Inventor, Bambu Labs

Designed, prototyped and 3D-printed a jack stand for my final design and documentation project.

This Project includes:

• Advanced CAD and 3D printing features
• Dimension comparisons between design and printed part
• Detachable handle system for convenient part replacement
• Cost and waste analysis

Tools: HTML, SCSS, JavaScript, PostCSS

A webiste that I co-built for a cookie company located in Addis Ababa, Ethiopia.

My main contributions include:

• Responsive design with a joyful and delicious feel.
• Alternative color theme switcher.
• On-scroll animations.
• Flip card effect for the product items.
• On-hover enlarged product gallery.
• About us page.
• Submittable contact form, and more.

Tools: C++, Python, GitHub Actions (CI/CD), Pandas, Matplotlib

Developed a physics-based simulation to model and validate the vertical flight dynamics of a drone.

Implemented a C++ simulation using Runge–Kutta integration to model motion under thrust, gravity, and drag. Defined simulation API through Drone.h, Integrator.h, and a shared State data structure.

Designed an automated CI/CD pipeline that builds and runs the simulation, verifies numerical results against analytical expectations, generates flight plots, and deploys validated results via GitHub Pages.

The framework is structured for future expansion to 2D and 3D dynamics, and GUI.

Tools: Python, Pandas, SQL, Numpy, Matpoltlib

As a Data Analyst, I collected data from various sources and identified trends for report on emerging skills.

This project includes:

• Data Collection using API and Web Scraping.
• Data Wrangling
• Exploratory Analysis
• Data Visualization
• Dashboards
• Presentation

Tools: HTML, CSS, React, JSX, OMDb API

A movie posters search engine that allows users to search for and explore movie posters.

This project includes:

• Collection of movies data using the OMDb API.
• Search input that maps over the OMDb API and fetches the movie component that matches your search.
• JSX component for a movie poster.
• Responsive CSS design.