Google DeepMinddm_control: Software and Tasks for Continuous Control
A collection of Python libraries and task suites for reinforcement learning agents in an articulated-body simulation, with convenient bindings, procedural model manipulation, and task authoring.
SoundCloudAnnouncing Twinagle: Twirp and Protobuf for Finagle
Combining Twirp and Protobuf with Finagle to enhance serialization protocols
OpenAIOpenAI API
Exploring the capabilities and applications of the OpenAI API
Jane StreetUsing ASCII waveforms to test hardware designs
Utilizing ASCII waveforms as an automated method to test the accuracy of hardware designs
Google DeepMindAcme: A new framework for distributed reinforcement learning
A framework for building readable, efficient, research-oriented RL algorithms that enables simple descriptions of RL agents at various scales of execution, including distributed agents.
OpenAILanguage models are few-shot learners
Understanding the potential of few-shot learning in language models
SoundCloudStretch Opportunities for Engineers
Exploring the concept of stretch opportunities for engineers and how they can help foster growth and skill improvement.
OpenAIAI and efficiency
Exploring the intersection of AI and efficiency in modern technology.
SoundCloudOpen Sourcing Intervene
Open sourcing the solution for fixing a weird looking upload page in the US by the web team at SoundCloud.
OpenAIJukebox
Exploring the latest advancements in audio streaming technology with Jukebox
Jane StreetChrome extensions: Finding the missing proof
Exploring the history and benefits of Chrome extensions for improving workflow and integrating with backend servers
OpenAIImproving verifiability in AI development
Techniques to enhance the verifiability of AI development processes