Syam sai santosh Bandi

I’m a Computer Science and Engineering student at Indraprastha Institute of Information Technology, Delhi, working as an Undergraduate Research Assistant at the MIDAS Lab. My focus is turning messy real-world data into reliable pipelines—especially for textual, multimedia-adjacent tasks.

At MIDAS Lab, I developed methods for annotating and processing written data to support machine learning tasks. I also collaborated on efficient data preprocessing pipelines and feature extraction techniques with my guide, Dr. Rajiv Ratn Shah.

In my projects, I build end-to-end systems and validate them rigorously. For Topological Signatures of Citation Cartels, I engineered a multimodal detection pipeline over a network of 500,000 academic papers and 4.87 million citation edges, using a Teacher-Student LLM framework with Gemma-3 and fine-tuned SciBERT to classify citation edges into 6 semantic intents. I then applied Louvain community detection and created a 6-metric Composite Cartel Index to flag anomalous communities, including network resilience simulations with a proof of zero structural degradation to the giant component.

I also enjoy high-performance engineering and applied security. For GPU-Accelerated Block Ciphers, I implemented GPU-accelerated versions of 12 symmetric block ciphers (up to 800x speedup) and developed a native CTR-mode CUDA kernel to fuse counter generation, encryption, and XOR on-device for a 14.2x throughput increase, with correctness verification against NIST FIPS-197. Separately, I built Plek, a room booking platform using Django, React.js, and OAuth 2.0, targeting 1000+ concurrent users with 99.9% uptime and GDPR compliance via RBAC.