GitHub - TheUnsolvedDev/TheUnsolvedDev

Explore the research laboratory · Website setup and content notes

PhD Student · Data Science & AI · IIT Madras
_{Building agents that learn, explore, and make better decisions.}

🧑‍🔬 About me

I'm Shuvrajeet, a PhD student in Data Science and Artificial Intelligence at IIT Madras, exploring how intelligent agents can learn through interaction.

My interests sit at the intersection of reinforcement learning, generative models, and decision-making. I enjoy taking an idea from equations to a working implementation—and investigating what happens when the experiments disagree with the intuition.

🔬 Research focus: Deep reinforcement learning and diffusion-based exploration.
🎯 Core question: How can learned generative priors help agents explore more effectively?
♟️ Building: Self-play agents, game-playing systems, and continuous-control experiments.
⚙️ Engineering: TensorFlow pipelines, vectorized environments, and multi-GPU training.
📚 Learning: Probabilistic machine learning, optimization, and GPU programming.
🛠️ Beyond research: MLOps, robotics, Linux, and animation.
🤝 Open to: Research discussions and collaborations around RL and generative models.

Less hype. More experiments. Let the ablations speak.

🧠 Research interests

Domain	What interests me
🎮 Deep reinforcement learning	Actor–critic methods, off-policy learning, policy optimization, and sample efficiency
🌊 Diffusion models for RL	Learning action priors and investigating their role in exploration
🧭 Exploration & uncertainty	Structured exploration, behavioral diversity, and probabilistic reasoning
♟️ Self-play & planning	Game-playing agents, Monte Carlo tree search, and learning through competition
🦿 Continuous control & robotics	Sequential decision-making in control tasks and embodied environments
⚡ Efficient learning systems	Vectorized rollouts, GPU computation, and reproducible training pipelines

🔬 Research spotlight · Diffusion as an Exploration Prior

Can a diffusion model learn to propose useful exploratory actions while an RL agent learns how to refine and use them?

I'm investigating diffusion models as a source of structured exploration for reinforcement learning.

🌊 Learn a prior: Model action patterns from collected experience.
🧭 Guide exploration: Investigate alternatives to unstructured action noise.
🎯 Connect to control: Study how proposals interact with a learned actor.
📊 Evaluate carefully: Compare against verified baselines such as SAC.
🧪 Test the mechanism: Measure sample efficiency, returns, diversity, and computational cost.

Status: An ongoing research direction; its benefits need to be established through controlled experiments.