flâneur

yiannisha.dev

yiannisha.dev · 244 words · saved by 1 readers

i work on robotics software, inference optimization, GPU-heavy ML systems, and research around continual learning, memory and (random nerd-snipe) neural combinatorial optimization for real-world operation research use cases. LIFE OBJECTIVE My life's RL objective is ship things that matter, over a long horizon. STARTUPS 3 / 0 failed startups / success stories PUBLICATIONS 2 / 2 NeurIPS, MLSys / first-author Robotics systems ML infrastructure CUDA, Rust, and performance work SOME OF MY WORK realized return so far. TOOLS FOR HUMANITY ML and systems work around the Orb stack, spanning robotics-adjacent software, deployment-heavy inference paths, and performance-sensitive infrastructure. PROJECT Current PROGRESSIVE ROBOTICS Designed and optimized custom reinforcement learning environments for palletizing, pushing CPU bottlenecks onto the GPU and making training loops materially faster. PROJECT 2025 - 2026 DELIAN Built the low-latency teleoperation stack for deployed systems, including strea

Yiannis Hadjiyianni ML Systems / Robotics / Performance i work on robotics software, inference optimization, GPU-heavy ML systems, and research around continual learning, memory and (random nerd-snipe) neural combinatorial optimization for real-world operation research use cases. Life Objective My life's RL objective is max⁡π Eπ[∑tγt1{shippedt} HumanUtilityt]\max_\pi \; \mathbb{E}_\pi \left[\sum_t \gamma^t \mathbf{1}\{\text{shipped}_t\}\,\text{HumanUtility}_t\right] ship things that matter, over a long horizon. Startups 3 / 0 failed startups / success stories Publications 2 / 2…

saved by

related reading