Summary
Skanda Koppula is a Senior Research Engineer at Google DeepMind with 12 years of experience building scalable multimodal and self-supervised systems for dense language and video understanding. He holds BS/MEng degrees from MIT and has blended academic rigor with product-focused engineering across roles at NVIDIA, Square, and ETH Zurich (as a Swiss Government/ Fulbright scholar). His research contributions span energy-efficient DNN inference (MICRO 2019) to production-oriented model training and multi-camera autonomy work, reflecting strengths at the intersection of systems, ML, and embedded platforms. As a co-founder of MIT Driverless he shipped high‑precision perception models (89% mAP YOLOv3) and explored transferable driving policies, revealing a practical bent for turning research into winning, deployed systems. Based in London, he brings rare cross-domain fluency—from DRAM-aware model tuning to large-scale multimodal architectures—making him effective at bridging low-level optimizations and high-level model design.
12 years of coding experience
4 years of employment as a software developer
Master's degree Computer Science Concentration: AI/ML, Master's degree Computer Science Concentration: AI/ML at Massachusetts Institute of Technology