Hello!

My name is Adithya. Nice to meet you!

I’m a machine learning engineer at Meta, where I work on recommendations for Facebook Marketplace, and a Georgia Tech CS grad (BS + MS), specializing in machine learning, systems, and algorithms. I have a passion for building things at the intersection of machine learning, systems, and data. I’m always interested in connecting with folks working in recommendation systems, ML infrastructure, distributed systems, biotech, or climate tech, so feel free to reach out if that’s you!

Experience

Professional

I’m currently a Machine Learning Engineer for Meta, where I work on recommendation systems for Facebook Marketplace. I’ve worked on training new ranking models and designing backend infrastructure for item recommendation diversity, quality, and user personalization, as well as for enabling socialization across Marketplace. If you’ve liked or commmented on Marketplace listings, or seen “which product do you like more” polls on Marketplace, that’s me!

I’m currently open to new opportunities! I’m particularly interested in roles in recommendation systems, ML infrastructure, distributed systems, and biotech. If that sounds like you or your company, feel free to reach out through my LinkedIn below.

I’ve also interned at a couple companies as a student. I’ve worked two summers at AWS as a Software Development Engineer Intern, where I worked on the generative AI capabilities for AWS AppStudio. Before that, I interned at ServiceNow as a Software Engineering Intern, where I worked on the MetricBase time-series database.

If you want to learn more, feel free to check out my resume and my LinkedIn.

Research

My research lies on deep learning inference on edge devices. More simply, it’s about trying to train and deploy neural networks that require as little computational power as possible to run on devices like laptops, smartphones, or even Raspberry Pis.

Most of my work is in post-training optimization for deep learning models. I even have a paper presented at NeurIPS 2024 about an algorithm I designed to improve the convergence of pruning-while-training algorithms for convolutional neural networks and transformer models. You can read the paper here.

In addition, I also have done work with others on improving the performance of foundational small language-vision models like LLaVA to get them to perform better on video captioning, annotation, and summarization — presented at the Workshop on Multimodal Generation and Retrieval at ACM MM 2024. You can read about that and look at the code over here.

Education

I graduated from the Georgia Institute of Technology with both a Bachelor’s and Master’s degree in Computer Science. My specializations are in Machine Learning and Computer Systems, but I also have a background in mathematics, algorithms, and theoretical computer science.

Projects

Outside of work and research, I’ve worked on a couple fun projects and experiments. Some highlights include a series of neural networks that attempt to fill out the March Madness bracket every year (check it out here), and a 2nd-place best overall project from hackGT X, that attempted to flag a voice as potentially showing signs of Parkinson’s Disease using a combination of audio processing and deep learning (check it out here).

You can find some of my other work at my GitHub.

Hobbies

Outside of work, I’m an avid sports fan. Though I mainly watch basketball (NBA, and men’s and women’s college, go Warriors and Yellow Jackets!), I do enjoy football (go 9ers), track and field, and recently picked up ice hockey as well (I’m an unfortunate Sharks fan). I’m also a big fan of role-playing video games, hiking, and skiing.

I also am a pretty big distance runner, having run a couple half-marathons, and plan to work up to a full marathon soon.

Thanks for stopping by!

Thanks for reading through everything! If there’s anything else you want to know, feel free to reach out through the links below. Have a great day!