Rahul Yedida

Resume / CV

I am qualified for the roles below. Note that each CV lists different sections and projects, based on the role. If you prefer, you can also download my master CV.

Academic

Most of my publications are in SE venues, since my research is at the intersection of deep learning and software engineering. My academic CV is below.

CV

Research Statement

Teaching Statement

Diversity Statement

ML Engineer / Research Scientist

My ML experience lies primarily in applied ML, although I have worked in theoretical research. I am comfortable using Keras, PyTorch and other common ML frameworks. My resume for ML Engineer, ML Research Scientist, and related positions is below.

CV

Software Developer

I have worked on multiple full-stack solutions using React/TypeScript and various database and server systems. My resume for software developer roles is below.

Download

Research Interests

AI for SE

I work at the intersection of artificial intelligence and software engineering. My work includes a novel SOTA hyper-parameter optimization method that requires 25% of the runtime and a novel oversampling method that improves defect prediction by up to 123%. My dissertation focused on better, faster deep learning for SE by using feedforward networks over static code features with theoretically-driven novel methods that outperform the prior state-of-the-art at a fraction of the cost.

Theory-Driven Deep Learning

I work on theoretical and applied deep learning, especially with loss functions. A lot of my applied deep learning work involves using recent results from the theoretical literature, bridging the gap between the two. Examples of this can be seen in my work on automated microservice partitioning and hyper-parameter optimization. Many of my core insights come from optimizing for better properties of the loss.

Recent Publications

Is Hyper-Parameter Optimization Different for Software Analytics?

Yedida, R., & Menzies, T.

IEEE Transactions on Software Engineering, 2025

(Re)use of Research Results (is Rampant)

Baldassarre, M. T., Ernst, N., Hermann, B., Menzies, T., & Yedida, R.

Communications of the ACM, 2023

How to Find Actionable Static Analysis Warnings: A Case Study with FindBugs

Yedida, R., Kang, H. J., Tu, K., Lo, D., & Menzies, T.

IEEE Transactions on Software Engineering, 2023

An Expert System for Redesigning Software for Cloud Applications

Yedida, R., Krishna, R., Kalia, A., Menzies, T., Xiao, J., & Vukovic, M.

Expert Systems with Applications, 2023

How to Improve Deep Learning for Software Analytics (a case study with code smell detection)

Yedida, R., Menzies, T.

IEEE/ACM 19th International Conference on Mining Software Repositories (MSR), 2022

Projects

Below is a sample of my projects. For a more comprehensive list of my notable projects, please view my master CV.

Zotero RAG System

A work-in-progress, I'm building a RAG (Retrieval-Augmented Generation) system for my growing Zotero library, which has gotten unmanageably large. The final vision is a system that can fetch references based on descriptions of papers and a QA system that is grounded in references. I'm writing this in Rust, partially to learn the language, but also to use a low-level language so I understand details of how things work without abstractions. As part of this project, I'm writing a simple PDF parser that extracts text and parses equations from academic PDFs.

Link: GitHub

Programmable Resumes

Not quite satisfied with the flexibility of existing resume solutions, I developed a syntax for modular resumes that are far more customizable and allow for generating multiple CVs in parallel. My CVs above are generated using this syntax.

Link: GitHub

pysh

A superset of Python that allows inline evaluation of Shell commands. I wrote the transpiler and a VS Code extension for syntax highlighting.

Link: GitHub

raise-utils

A Python package written to centralize the implementations of our lab's algorithms. Currently at 50k downloads.

Link: PyPI and Code

Web Dev Mini-Projects

While learning web development at FreeCodeCamp, I developed several small projects using the MERN stack, including a URL shortener, a voting application, a rogue-like dungeon crawler game, a Simon game, and several visualization projects using D3.js

Link: CodePen

JournalBear

A cross-platform journal application built with Electron, encrypted by AES-256.

Links: Code and Software

Human Activity Data Project

This is a fun side project I worked on. For around 10 months, I collected data on every single activity I do, with a total of 30 activity categories. I then used this data to analyze my most productive hours and built a 2-layer LSTM to predict my next activity based on the previous five.

Links: Time Meter (Android) | ATracker (iOS)

Machine Learning Blog

A blog to teach machine learning to beginners in a way I believe that I would have found helpful when I had started, making sure it has the mathematical rigor, but also an intuitive explanation, and Python code.

Link: Blog.

Personal Interests

Coffee

I am a coffee enthusiast. Currently, I brew my pour overs using the Orea v3, and my espresso using the Breville Barista Express and a Sette 270Wi. My favorite coffee so far was a Panama Geisha from Black & White Coffee Roasters.

Taylor Swift

I've been a big fan of Taylor Swift since 2018! My favorite era is the Lover era, although I currently have evermore on repeat. My current favorite songs (in no particular order) are Hits Different,Cruel Summer (Live), Nothing New (Taylor' Version) (From The Vault),I Can See You (Taylor' Version) (From The Vault), You're On Your Own, Kid,and gold rush, although this list frequently changes. Yes, I'm excited for TTPD!

Digital Privacy

I am a big advocate for digital privacy and annually donate to organizations that actively fight for users' rights online.

Contact

Direct contact

E-mailPreferred