Akshada Devkar

I’m a final-year B.Tech AI & Data Science student with production ML experience from my internship at Reliance Jio. I built a GNN-based Network Digital Twin using Graph Attention Networks (GAT) to model live telecom topology, enabling failure simulation and anomaly propagation analysis before production deployment.

I also processed millions of daily network log rows with Python (Pandas, NumPy) to create anomaly detection pipelines that surfaced failure patterns before manual escalation, using techniques like K-Means for customer segmentation. Alongside this, I researched LLaMA architecture internals (RoPE, RMSNorm, grouped-query attention), and I enjoy turning research into shippable systems through projects like telecom GNN anomaly detection dashboards and LLM-based summarization.