Curriculum Vitae

• Built a ConvLSTM lightning prediction system end-to-end in PyTorch, achieving 92% accuracy on live meteorological data
• Deployed to active meteorologist workflows at NRSC (production system with real users)
• Reduced training time by 35% through optimisation on large-scale simulation datasets using a VAE-based approach
• Built the React dashboard used by meteorologists to monitor model outputs in real time

• Built stock price directional models achieving 84% accuracy
• Developed data pipelines and Tableau dashboards for investment teams

• Shipped four Java web applications from scratch to production using JSP and Tomcat

• Built a basic Distributed Key-Value Store implementation in Go using Raft consensus
• Developed Career Radar, a live job-fit analyser that matches a GitHub profile against current market job listings

• Parallel programming on ARCHER2 (UK National Supercomputer) using MPI, OpenMP, and CUDA
• Dissertation: Early ML-based detection of memory leaks in containerised environments (in progress)
• Supervised by Ben Carpenter, EPCC Applications Developer; using cAdvisor, Prometheus, eBPF, TimescaleDB, Grafana

• Final year project: GaN-Enhanced Model (semiconductor device simulation and ML)
• ISRO internship completed during final year

ISRO Lightning Prediction System

ConvLSTM model in PyTorch for lightning strike prediction. 92% accuracy on live WRF data, deployed to active NRSC meteorologist workflows.

MSc Dissertation: Memory Leak Detection

ML-based early warning system for memory leaks in containerised environments. Uses eBPF, Prometheus, TimescaleDB, and Isolation Forest / LSTM detection models.

Career Radar: Job-Fit Analyser

Live tool that matches a GitHub profile against current job listings, outputting fit scores and ranked skill gap analysis.