What I Focus On

• Cross-subject biosignal decoding — EEG and wrist surface EMG models that work on new people with little or no calibration. I’m currently an independent entrant in the NeurIPS 2026 EEG/EMG Foundation Challenge (Track 2, EEG decoding), with a covariance-based decoder (Riemannian tangent space with logistic regression) evaluated on held-out participants.
• Federated learning — personalising shared models without pooling each person’s data. At VIT Chennai I led HabitatFL, a federated framework for classifying wildlife audio and images, built with Flower and Docker on the university’s NVIDIA DGX; its edge models (under 2 MB) reached 87.65% accuracy on audio and 69.92% on images.
• Leakage-free evaluation — subject-wise splits that rule out data leakage.

Recent Highlights

• Two peer-reviewed papers: JIGS in Engineering Research Express (2025), a parallelised ECDH and AES-128 secure-messaging pipeline verified in ProVerif, and Siren in Springer LNNS (2025), a deception-based cyber defence
• Application Engineer at Accenture (Oct 2025 – Oct 2026): built Salesforce Life Sciences Cloud features for a global pharmaceutical client, and evaluated GitHub Copilot on real user stories (over 80% average time savings), which fed into the licence decision
• Founder and writer of GreyLattice (since Feb 2025): around ten posts on brain–computer interfaces and computational neuroscience
• GPU-accelerated financial news sentiment with Kafka, FinBERT and RAPIDS cuDF: end-to-end latency cut from about 8 minutes to 10 seconds
• Talk: introduction to CUDA programming, VIT Chennai (50+ attendees)
• B.Tech. in Computer Science and Engineering, VIT Chennai (2021–2025), CGPA 8.69/10

Open To

• Research collaborations in EEG/EMG decoding, federated learning, or computational neuroscience
• Internships with labs/companies advancing AI systems, BCI, or edge computing
• Projects at the boundary of engineering and research