Software Engineer · Backend & Infrastructure · Security · Agentic AI
Building things that don't break under pressure
I'm a backend-focused software engineer who's most comfortable at the intersection of distributed systems, infrastructure, and security. I've shipped production systems across self-hosted Linux deployments and GIS applications — and I care deeply about the stuff that holds it all together: reliability, threat modelling, and clean architecture.
- 🎓 B.S. CS @ LUMS, Batch '26
- 🔭 Currently building ContinuumAI — an autonomous agentic data scientist using MCP
- 🌍 Previously contracted with National University of Singapore on a GIS flood research platform
- 🛡️ Interested in engineering production systems end-to-end — from architecture and reliability to security hardening
Languages
Backend & AI
Data & Infrastructure
🤖 ContinuumAI — Agentic AI Data Scientist (Senior Capstone)
Autonomous agent using Model Context Protocol (MCP) that reasons over complex datasets in real time.
- Exposes 5+ tools to a Python inference engine for multi-step decision-making
- Built a Parquet-to-SQL ETL pipeline + Chart Spec engine for structured data querying
- Hardened via STRIDE threat modelling & SonarQube — resolved 15+ security issues
🌊 HiPer-FedKD — Federated Learning Framework
Novel FL framework tackling data heterogeneity in distributed edge-AI environments.
- +16% accuracy gain via bidirectional knowledge distillation
- Benchmarked on high-performance Linux clusters using PyTorch
🗄️ StayOpsApp — High-Availability DB on Kubernetes
HA SQL Server cluster targeting 99.9% uptime on Kubernetes.
- Persistent volumes, network policies, and YAML-driven CI deployments
- Tuned 1M+ row database via partitioning and execution plan analysis
Distributed key-value store with full Raft consensus across 5 nodes.
- Sustains strong consistency under 2 simultaneous node failures
- Resolved 10+ race conditions; validated under adversarial network partition scenarios
Open to backend, infrastructure, or security engineering roles — remote or on-site.
Let's build something reliable.