00Profile
Building systems that understand, adapt, and prevent. So no human brain has to suffer.
View work
01About
AI Engineer building at the intersection of deep learning, reinforcement learning, and neuroscience.
MS in Artificial Intelligence from the University at Buffalo. BTech in Computer Engineering (AI/ML) from Silver Oak University.
Previously: IEEE Computer Society & Signal Processing Society Chairperson at Silver Oak University.
“First-principles algorithms over surface-level abstractions. AI that augments existing systems rather than replacing them.”
01.1Expertise
EEG Conformer, Mamba SSM, VAE, diffusion models for sepsis treatment
AWR-based offline RL agent, decision optimization under sparse rewards
Production model serving, latency optimization, quantization pipelines
LLM rationale generation, multi-step reasoning systems
Medical imaging analysis, real-time detection systems
CI/CD for ML, experiment tracking, model monitoring at scale
02Research
Academic contributions at the intersection of deep learning, reinforcement learning, and neuroscience.
HDBSCAN + VAE + diffusion models + AWR-based RL agent + LLM rationale generation. Achieved 83% treatment accuracy on MIMIC-III and eICU datasets.
Co-authored with Punit Kumar, Vaibhav Saran, Nitin Kulkarni, Alina Vereshchaka.
Mamba SSM + CNN encoder with three-component loss (contrastive + forecast + biomarker alignment). Near-zero forecast MSE (~0.002) on EEG temporal dynamics.
Co-authored with Rajvi Zala.
Introduced the 81-dimensional Spectral-DASI feature token and Dynamic Alpha Suppression Index (DASI) for binary stress classification. EEG Conformer and EEGMamba backends on SEED-VII.
03Focus Areas
Where research and shipping overlap.
End-to-end intelligent systems that learn, adapt, and optimize in real time. From reinforcement learning agents to production-grade computer vision and MLOps pipelines.
Real-time decision engines for environments with sparse rewards and delayed feedback. Turning uncertainty into action.
Real-time neurological signal processing for proactive detection and prevention. Because no human brain should have to suffer.
Tools, frameworks, and research code shared with the community. Building in the open wherever possible.
04Blog
Long-form writing on reinforcement learning, AI, and the future of intelligence.
Breaking down artificial general intelligence into concepts anyone can understand.
Should AI be tailored to the user or to the task? A reinforcement learning take on the question.
05Contact
Interested in collaboration, research, or just want to talk about AI and neuroscience? I'd love to hear from you.
pateldivyesh2309@gmail.com