Tejan Karmali
Graduate Student Researcher
Palo Alto, California, United States
Summary
🤩
Rockstar🎓
Top SchoolTejan Karmali is a PhD student and graduate researcher at Stanford with nine years of industry and research experience building deep learning systems for generative modeling, reinforcement learning, and Earth observation. He has driven production-facing projects at Google (multi-frame super-resolution and super-resolved segmentation for satellite imagery) and led image-to-3D reconstruction efforts at Avataar, demonstrating both applied research impact and system-level engineering. An active open-source contributor to Flux.jl and the Flux model-zoo, he implemented optimizers like NADAM and RL algorithms (DDPG, DQN), showing a rare fluency in low-level ML tooling as well as high-level model design. Comfortable moving between Julia libraries, Python ecosystems, and large reconstruction/diffusion models, he focuses on exploration and sample-efficient methods in reinforcement learning for his doctoral work. Tejan pairs strong experimental results (e.g., substantial InstanceIoU gains on super-resolved segmentation) with practical engineering—building data pipelines, refactoring core libraries, and shipping reproducible research.
9 years of coding experience
2 years of employment as a software developer
Master of Technology (Research) Computational and Data Sciences, Master of Technology (Research) Computational and Data Sciences at Indian Institute of Science (IISc)Bachelor of Technology (B.Tech.) Computer Science and Engineering, Bachelor of Technology (B.Tech.) Computer Science and Engineering at National Institute of Technology, GoaHSSC Science, HSSC Science at Vidya Vikas AcademySSC, SSC at Mahila & Nutan English High SchoolDoctor of Philosophy - PhD Computer Science, Doctor of Philosophy - PhD Computer Science at Stanford University
English, Hindi, Konkani, Marathi, French
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (17)
artificial-neural-networks10
optimizers10
machine-learning10
optimizer10
reinforcement-learning10
deeplearning-ai10
deep-learning10
ddpg10
flux10
neural-network10
fluxor10
julia10
dqn9
python8
zoo8
Programming languages (7)
Github contributions (5)
FluxML/model-zoo
Jan 2019 - Jun 2019
Please do not feed the models
Role in this project:
ML Engineer
ML EngineerContributions:47 commits, 12 PRs, 3 pushes in 5 months
Contributions summary:Tejan contributed significantly to a reinforcement learning project within the repository. They implemented a DDPG (Deep Deterministic Policy Gradient) algorithm for a Trebuchet simulation, including defining actor and critic models and training functions. Further contributions included the creation of a DQN (Deep Q-Network) example on CartPole with Gym, demonstrating an understanding of various reinforcement learning techniques and their application. They also refactored existing code, updated dependencies, and fixed various errors.
deep-learningmachine-learningfluxfeedjulia
FluxML/Flux.jl
Mar 2018 - May 2019
Relax! Flux is the ML library that doesn't make you tensor
Role in this project:ML Engineer
ML EngineerContributions:36 commits, 13 PRs, 1 push in 1 year 1 month
Contributions summary:Tejan implemented and refined various optimization algorithms for machine learning models within the Flux.jl library. Their contributions included the addition of the NADAM optimizer, fixing update rules, moving epsilon values, and incorporating improvements to existing optimizers like Adam and RMSProp. Additionally, the user made adjustments to the Conv and ConvTranspose layers, demonstrating a focus on the core components of neural network development within this machine learning framework.
ml-librarythe-human-braindata-sciencedeep-learningneural-networks
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.