A
Aravind Srinivas
About
Detail
Cofounder, President, CEO
San Francisco, California, United States
Résumé
Jobs
verified_user
0% verified
- AAngel InvestorAravind SrinivasJan 2023 - Current (3 years 6 months)Cursor, Eleven Labs, Suno, Mistral, Cognition, Lovable, Fal, LiveKit
Cofounder, President, CEOPerplexityAug 2022 - Current (3 years 11 months)Fastest way to get answers to any question- Research ScientistOpenAISep 2021 - Aug 2022 (1 year)Research on language and diffusion generative models.
- RResearch InternMay 2020 - Apr 2021 (1 year)Multimodal Transformers: Bottleneck Transformers, Local Self-Attention
- Research InternDeepMindMay 2019 - Sep 2019 (5 months)Large-Scale Contrastive Learning
- Research InternOpenAIMay 2018 - Aug 2018 (4 months)Research on Policy Gradient Algorithms.
Education
verified_user
0% verified
- Doctor of Philosophy - PhD, Computer ScienceUNIVERSITY OF CALIFORNIA BERKELEYResearch: Contrastive Learning for Computer Vision (CPCv2), Reinforcement Learning (CURL). Transformers for Image Generation (Flow++), Image Recognition (BoTNet), Video Generation (VideoGPT), Reinforcement Learning (Decision Transformer). Teaching: Deep Unsupervised Learning (Spring 20' and 21').
- Dual Degree (B.Tech and M.Tech, Electrical EngineeringIndian Institute of Technology MadrasResearch: Deep Reinforcement Learning: Dynamic Action Repeats for Hierarchical RL (AAAI 17', ICLR 17'), Transfer Learning (ICLR 17'). Teaching: Reinforcement Learning (Spring 17'), Advanced Reinforcement Learning (Spring 17').
Awards
verified_user
0% verified
- IViterbi India FellowshipIndoUS science and technology forumMay 2015https://www.iusstf.org/assets/sitesfile/image/download_links/pdf/pdf_2068068638.pdf
- NIndian National Mathematics Olympiad Merit AwardeeNational Board of Higher Mathematics India
- DKishore Vaigyanik Protsahan YojanaDepartment of Science and Technology Government of India
- GNational Talent Search ScholarshipGovernment of India
Publications
verified_user
0% verified
- rHierarchical Reinforcement Learning using Spatio-Temporal Abstractions and Deep Neural Networksrd International Conference on Machine Learning ICMLJun 2016This paper introduces an automated skill acquisition framework in reinforcement learning which involves identifying a hierarchical description of the given task in terms of abstract states and extended actions between abstract states. Identifying such structures present in the task provides ways to simplify and speed up reinforcement learning algorithms. These structures also help to generalise such algorithms over multiple tasks without relearning policies from scratch. We use ideas from dynamical systems to find metastable regions in the state space and associate them with abstract states. The spectral clustering algorithm PCCA+ is used to identify suitable abstractions aligned to the underlying structure. Skills are defined in terms of t