Software Engineer 2 on Cisco's AI Acceleration Team, Bengaluru. Research Associate at SJSU's MiCoSyS Lab with Dr. Saptarshi Sengupta, working on temporal graph learning.

Research site · Portfolio · Google Scholar · LinkedIn · gauthamprabhu9@gmail.com

Most of my work is about getting dependable results out of models under a fixed budget. The budget changes: qubits, milliseconds, trainable parameters. The question does not.

Temporal graphs and language models (SJSU, ongoing). TGaLoRA pairs a memory-based TGN, pre-trained on the Enron email graph (~798K timestamped edges, 42.7K nodes), with a gated LoRA adapter inside GPT-Neo-125M for temporal link prediction. The adapter adds 0.24% parameters. It underperformed a strong structural baseline, and finding out why turned the project into a reproduction and ablation study. Workshop paper under review at GFM@ICML 2026.

Quantum ML for cardiac diagnosis (Manipal, 2022 to 2023). QuCardio benchmarks quantum kernel methods (QSVC, Pegasos QSVC) against a quanvolutional neural network for ECG-based cardiovascular disease detection. The QNN reached 97.31% accuracy. Funded by a $27,000 MeitY / AWS Quantum Computing Applications Lab grant. Published in IEEE Access, 60+ citations.

Retinal imaging (IIT Kharagpur, summer 2023). With Dr. Subhamoy Mandal at the Medical Informatics Lab: a cross-attention vision transformer that fuses thermal and OCT structural maps to identify age-related macular degeneration.

Earlier work (Manipal). Attention and residual U-Nets for retinal vessel segmentation, with Diffie-Hellman + AES-256 encryption so patient images stay private (EyeEncrypt, with Dr. Srikanth Prabhu). BERT and XLNet classifiers for vaccine misinformation at 94.84% accuracy, plus a Reddit user-association graph with community detection to find high-leverage accounts (with Dr. Nisha P. Shetty).

Cisco, January 2024 to now. Technical Intern, then Software Engineer 1 (August 2024), then Software Engineer 2 (October 2025).

I build agentic AI systems for Cisco's supply chain: automated test triage, root-cause analysis, and retrieval over internal knowledge through MCP servers. 35% of support cases now resolve without an engineer. 70% of previously undiagnosed failures now get a root cause. MTTR is down 40%, and the team saves roughly 10,000 engineer-hours a quarter. I also shipped a computer vision check for edge gaps in connectors on the factory line, and cut an internal deployment from 6 hours to under 2 minutes with Jenkins and Docker.

1. QuCardio: Application of Quantum Machine Learning for Detection of Cardiovascular Diseases. S. Prabhu, S. Gupta, G. M. Prabhu, A. V. Dhanuka, K. V. Bhat. IEEE Access, vol. 11, 2023. doi
2. Addressing Vaccine Misinformation on Social Media by Leveraging Transformers and User Association Dynamics. C. Rao, G. M. Prabhu, A. R. Kumar, S. Gupta, N. P. Shetty. Procedia Computer Science (ICMLDE 2023), vol. 235, 2024. doi
3. VIKAS: A Multimodal Framework to Aid in Effective Disaster Management. G. M. Prabhu, T. Gupta, M. V. Srujan, A. R. Soumya, A. Palorkar, A. Chowdhury. ATIS 2022, Springer CCIS, vol. 1804, 2023. doi
4. EyeEncrypt: A Cyber-Secured Framework for Retinal Image Segmentation. G. Hegde, S. Gupta, G. M. Prabhu, S. V. Bhandary. ATIS 2022, Springer CCIS, vol. 1804, 2023. doi
5. A Systematic Review of Deep Learning Approaches for Vessel Segmentation in Retinal Fundus Images. G. Hegde, S. Prabhu, S. Gupta, G. M. Prabhu, A. Palorkar, M. V. Srujan, S. V. Bhandary. Journal of Physics: Conference Series, vol. 2571, 2023. doi
6. Exploring Design Optimisation Techniques of a Radio Telescope using Fixed Costing Constraints. A. A. Iyer, G. M. Prabhu, T. Gupta, S. Deshmukh, R. Rivankar. Journal of Physics: Conference Series, vol. 2571, 2023. doi

Talks: SatelTensor, satellite data exploration with tensor decomposition (TCML Workshop, IISc, 2023). Multimodal ocular diagnostics (Digital Health Symposium, IIT Kharagpur, 2024).

• proctor: reads Claude Code session transcripts and puts a dollar figure on wasted tokens. Standard library only, runs locally.
• SystemDesignCode: open-source practice platform for machine coding and LLD rounds, with sandboxed execution and an AI design reviewer.
• VIKAS: disaster-response pipeline over social media text and images. I led the team. Smart India Hackathon 2022 Grand Finalist, presented to the NDRF.
• Quantum Ecosystem for Medical Imaging: the 2022 hackathon prototype that came before QuCardio.
• misc-learning-projects: small things I built while learning, 2020 to 2023.

• Winner, Cisco CTF Offensive Summit
• 2nd Runner-Up, Cisco Intern Case Study Competition (NLP supplier-name standardisation)
• Grand Finalist, Smart India Hackathon 2022 (team lead)
• Grand Finalist, Quantum Science and Technology Hackathon 2022 (top 16 of 1,600+ teams)
• NTSE State Scholar (rank 21 of 151,000+)

Co-founder and technical head of Project Kalpana, an 8-person student radio astronomy team that won a $13,000 R&D grant. General Secretary and Treasurer of ACM Manipal. Technical Head of the Astronomy Club of Manipal.

B.Tech in Computer Science, Manipal Institute of Technology, 2024. CGPA 8.91/10, minor in Big Data Analytics.