Asad Jeewa
Lecturer at University of Cape Town
Western Cape, South Africa
Summary
👤
Senior🎓
Top SchoolAsad Jeewa is a Computer Science lecturer and PhD candidate specializing in Multi-Objective and Multi-Agent Reinforcement Learning, with nine years of experience bridging academia and industry. He has taught across South African universities while contributing to applied research at InstaDeep (MAV A/JAX codebase) and IBM Research Africa (Clinical NLP), bringing both production-focused engineering and experimental rigour to his projects. His MSc (cum laude) explored generalisation and reward engineering in sparse-reward RL, and he holds practical software credentials including Microsoft C# certification and simulation work for heavy machinery. Comfortable moving between lecturing, research engineering, and open-source contributions, he has hands-on experience implementing data storage and executor changes in large RL codebases, reflecting a knack for scalable RL systems engineering.
9 years of coding experience
2 years of employment as a software developer
High School, High School at Durban High SchoolDoctor of Philosophy - PhD Computer Science, Doctor of Philosophy - PhD Computer Science at University of KwaZulu-Natal
English, Afrikaans, Zulu, Urdu
Github Skills (6)
machine-learning10
jax10
multi-agent-reinforcement-learning10
python10
reverse10
reinforcement-learning9
Programming languages (3)
Github contributions (5)
instadeepai/Mava
Oct 2021 - Sep 2022
🦁 A research-friendly codebase for fast experimentation of multi-agent reinforcement learning in JAX
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:431 reviews, 229 commits, 42 PRs in 11 months
Contributions summary:Asad contributed to the implementation of a Reverb Parallel Adder for multi-agent reinforcement learning in JAX. The user's commits focused on modifying the codebase to incorporate Reverb for data storage and retrieval, with specific attention to the `MADQNFeedForwardExecutor` and `MADQNRecurrentExecutor` classes. The user also made changes to address shared weights across agent types and removed redundant code related to value networks, showcasing their involvement in core model architecture and data handling.
multi-agent-systemsmultiagentagentreinforcement-learningreinforcement-learning-agent
AsadJeewa/meltingpot
Jul 2023 - Dec 2024
A suite of test scenarios for multi-agent reinforcement learning.
Contributions:48 pushes, 2 branches in 1 year 5 months
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.