About Me
Hello! My name is Oriel Savir, and this is my website. I'm a student at Johns Hopkins University studying CS & Applied Maths (Financial Mathematics). I'm a curious mind with a passion for technology, math, and innovation. Throughout my time in college, I've gained experience in software development, deep learning research, and open source tech. In addition, I love using my math skills to solve complex problems, especially in deep learning, data science, computational biology, and quantitative finance! Whether it's engineering complex software or pushing the boundaries of deep learning, I'm always eager to learn more and collaborate on exciting projects with passionate, talented people. Feel free to reach out, would love to chat!
Skills
- • Python, C++, C, TypeScript, JavaScript, Java
- • React, Next.js, Express
- • PostgreSQL, MySQL, Snowflake, Apache Spark
- • AWS (S3, Lambda, EMR), Databricks, Docker
- • PyTorch, NumPy, pandas, SciPy
Interests
- • Deep Learning
- • AI-enabled technology
- • Full-Stack Development
- • Open Source
- • Dev Tools
Some of my projects
Low-Latency Portfolio Return Analyzer
A C++ tool for real-time portfolio return analysis, leveraging non-linear optimization and efficient data structures for quantitative finance applications.
EndoVis Segmentation Challenge
A deep learning-based U-Net model for surgical tool segmentation in endoscopic images, optimized for robustness against occlusions and real-world medical imaging tasks.
Medslate
An AI-powered web app that transcribes medical appointments and generates patient-friendly summaries and follow-up questions using OpenAI and AWS.
My work experience
Software Engineering Intern
Capital One
Developed a Python SDK with a programmatic API for feature store microservices, enabling ML and big data teams to create, manage, and query feature stores at scale. Implemented Capital One’s first feature store within the credit decisioning data pipeline. Developed a modular Snowflake API interfacing on AWS EMR and Databricks
Software Engineering Intern
JHU COLLAB
Developed and deployed a full-stack web application in Next.js (TypeScript) for extracting tables from unstructured documents, integrating a multi-branch CNN with 98% accuracy. Built a RESTful API with serverless endpoints to integrate with AWS S3 and AWS SageMaker, supporting 1000 concurrent tasks and decreasing inference time by 40%.
Deep Learning Researcher
JHU Department of Computer Science
Author and researcher on Normalization-Equivariant Learned Proximal Networks, a novel architecture for inverse problems with state-of-the-art noise signal robustness. Implemented CNNs using PyTorch, achieving an over 100% robustness increase on benchmarks.
Senior Teaching Assistant, Deep Learning (CS 482/682)
Johns Hopkins University
Support 150+ students in a graduate-level deep learning course covering supervised and unsupervised learning, neural architectures, optimization, and novel applications.
Computational Biophysics Research Assistant
JHU Department of Biophysics
Built computational models involving stochastic reaction-diffusion continuum models of biological systems using Python. Developed C++ simulations to study the impact of cargo binding on clathrin complex longevity.
Software Engineer
Delineo Disease Modeling Group
Led a development team modeling disease spread mechanisms using Python and Next.js. Developed machine learning modules integrating the SynthPops library to simulate social contact networks across five major cities. Added support for MongoDB to store simulation results, built a frontend to display results interactively, allowing for deeper data analysis insights.
Wanna talk about something? Shoot me a message!
Let's Connect
I'm always open to discussing new projects, ideas, or opportunities. Feel free to reach out here or through my socials!
Personal Email
orisavir at gmail dot comSchool Email
osavir1 at jhu dot eduLocation
New York, NY | Baltimore, MD