Shreyas Bachiraju

I’m a Master’s student in Machine Learning at Carnegie Mellon University, where I focus on efficient and reliable machine learning systems. I completed my B.S. in Informatics at Arizona State University, where I was fortunate to conduct research under the supervision of Prof. Hua Wei and Prof. Krishnendu Chakrabarty.

Broadly, I’m interested in making large language models more capable, more efficient, safe, and more sustainable. My research interests include:

• Sustainable AI — energy-efficient LLMs, and AI applied to sustainability problems
• Efficient LLMs — inference, serving systems, and model compression
• AI Safety — alignment, robustness, and reliable model behavior
• LLM Agents — reliable multi-step reasoning and tool use

The thread running through all of it is AI for good: building systems that are efficient, safe, and genuinely useful to the people and communities they serve. That means doing more with less — less compute, less energy, less waste — while making sure the models we build behave the way we intend. Recent projects include compressing 100K-step reinforcement learning runs into tiny synthetic datasets, profiling energy bottlenecks in Mixture-of-Experts language models, generative AI for transportation planning, and optimizing neural networks for edge devices like the Jetson Nano.

I was named a Fulton Outstanding Graduate by ASU’s Ira A. Fulton Schools of Engineering, won 1st Place at HackHarvard 2024, and received an Honorable Mention for the CRA Outstanding Undergraduate Researcher Award (2025–26). Beyond research, I co-founded CS + Social Good at ASU to build technical tools for NGOs. Outside of academia, I’m an avid football (soccer) fan and a music producer.

I believe technical excellence and thoughtful consideration of AI’s societal impact should go hand in hand. Feel free to reach out over email.