An AI researcher, Professor and Consultant.
Some of the projects I have worked on
​I had the privilege to work on several impactful projects on advancing AI algorithms, some of the public projects are listed here:
​
-
Foundation models and improvements over LLMs: Griffin and Hawk (2024), SDTT (2024), Hyperbolic attention networks (2018), RNN Enc-Dec (2014), GRU (2014), GORU (2019), DT-RNN (2013), LRU (2022), PAG (2017), Recurrent Batch Normalization.
-
LLM agents, reasoning, alignment, and post-training: Fleet of Agents (2024), LowRank and BlockDense (2024), ReST (2023), A2T (2024), DNTM (2015), Tardis (2016), SIKeD (2024).
​
-
RL and imitation learning: R2D3 (2019), BVE (2020), CRR (2020), Alphastar (2019), Offline-Actor Critic (2022), Muzero Supervised (2022).
-
Benchmarks, Datasets, and Evaluations: RL Unplugged (2020), LLM Self Recognition (2024), Swiss political alignment dataset (2024), Alphastar Unplugged (2022).
-
Software and Benchmarks: Theano (2014), Groundhog (2014) - First seq-to-seq open source framework (2013), Pylearn (2021), Acme (2022), RL Unplugged (2020), Hard-Eight (2021), Stable Diffusion Lens (2024).