Krishna Teja Chitty-Venkata
Senior Machine Learning Research Engineer at Red Hat
Benchmarking and optimizing LLM inference performance across diverse hardware accelerators including NVIDIA H100, AMD MI300X, and Gaudi2.
Intro
I am Krishna Teja Chitty-Venkata, a Senior Machine Learning Research Engineer at Red Hat. My career has been defined by a deep commitment to making AI more efficient, moving from a PhD at Iowa State to postdoctoral research at Argonne National Laboratory. I specialize in the intersection of High-Performance Computing and Machine Learning, specifically focusing on LLM inference optimization, quantization, and hardware-software co-design. I am passionate about replacing manual trial-and-error in network design with automated methods like Neural Architecture Search. Currently, I am focused on scaling the deployment of LLMs and am actively looking to connect with experts in inference and quantization to join our team at Red Hat. Whether you are interested in technical collaboration, research insights, or exploring opportunities in ML research, I am always open to meaningful professional dialogue.
Networking
What I can offer
- ›Expertise in quantization and LLM compression
- ›Insights into AI accelerator performance (NVIDIA, AMD, Intel)
- ›Collaboration on high-performance computing and ML research
- ›Hiring opportunities within the Red Hat ML Research team
What I'm looking for
- ›Experts in inference and quantization for recruitment at Red Hat
- ›Professional collaborations or invitations for short talks
- ›Students with strong programming and AI skills for mentorship
Best fit for
Focus
Current interests
Core competencies
Background
Career
Progressed from research internships at AMD and Intel to a PhD at Iowa State, followed by a Postdoctoral residency at Argonne National Laboratory before joining Red Hat as a Senior ML Research Engineer.
Education
PhD in Computer Engineering from Iowa State University (2017–2023); Bachelor of Engineering in Electronics and Communications Engineering from Osmania University (2013–2017).
Achievements
- ›Received the Outstanding Postdoctoral Performance Award from Argonne National Laboratory
- ›Authored a comprehensive survey on optimizing Transformer inference reviewing 250+ papers
- ›Developed LLM-Inference-Bench for benchmarking models across NVIDIA, AMD, and Gaudi hardware
- ›Received the Research Excellence Award from Iowa State University Graduate College
Opinions
- Understanding and improving inference efficiency is critical for scaling the deployment of LLMs across industries.
- Manual neural network design is inefficient; automated methods like NAS should replace trial-and-error evaluations.
- Open research and sharing code repositories are essential to foster community knowledge.
Personality
Communication style
Professional, academic, and collaborative, often using structured lists to summarize technical highlights.
Formality — 8/10
Vocabulary