At IIT Kharagpur, I designed DCAF-Net, a cross-attention network for real-time vibration classification. It achieved 98.2% mean ternary accuracy across four unseen geometries, and I deployed a CPU-only pipeline with under 100 ms latency.
I built Gnanam, an offline-first mobile AI tutor that runs quantized language models on-device with no network traffic. It includes on-device textbook OCR and local data storage for privacy.
On A.U.R.A., I built a multi-agent analytics pipeline that turns raw CSVs into interactive HTML dashboards while keeping raw data in DuckDB. I also worked on MediMitra & Rural Med AI / HAI, combining regional-language healthcare support with edge computer vision for rural screening.
I'm studying Computer Science and Engineering (Data Science) at MVGR College of Engineering, where my focus includes data science and generative AI. I'm preparing for GATE 2027 (DA) and am a national-level chess player.
