Open to opportunities

Syed Daniyal

@syeddaniyal1

Message

I'm a computer science researcher building deep learning systems for climate modeling and medical imaging.

Pakistan

Message

What I'm looking for

I'm looking to apply deep learning research and full-stack implementation skills to challenging AI systems, particularly in medical imaging, climate modeling, and generative AI.

I'm researching climate-data super-resolution at LUMS with Dr. Zubair Khalid, where I implemented and trained BLISSNet in PyTorch and built an ERA5 data pipeline for regional surface-temperature reconstruction. My cross-region evaluation found strong in-domain performance but transfer degradation tied to SIREN coordinate-domain dependence, informing the model's next iteration.

I also built ARIA, a neuro-symbolic radiology report-generation pipeline combining SegResNet, radiomics, Neo4j clinical rules, and constrained LLM generation. Using DeepSeek-R1, the system achieved a 0.718 GREEN score, improving accuracy by 105.1% and reducing hallucinations by 66.2% against the strongest unconstrained VLM baseline.

Experience

Work history, roles, and key accomplishments

Current

Undergraduate Research Assistant

Current

Lahore University of Management Sciences (LUMS)

Jun 2026 - Present (2 months)

Researching super-resolution of spatiotemporal climate data, focusing on reconstructing high-resolution surface temperature fields from sparse observations. Implemented and trained BLISSNet, a DeepONet-style neural operator, and evaluated cross-region generalization.

PyTorch Deep Learning Neural Operators NetCDF Attention U Net