Selected Projects
A selection of applied AI/ML systems I have designed and built, from national-scale LLM-powered cybersecurity infrastructure to agentic financial analysis pipelines, on-premise speech intelligence, and enterprise RAG. Each project combines research-level methodology with production-grade engineering.
Research Portfolio (This Site)
This site. A Next.js research portfolio with custom Research and Collaborations sections, canvas-based particle-network background (neural-network aesthetic), Clarity-inspired typographic design (EB Garamond + Inter), Journal/Conference publication filtering, a full writing section with featured-papers sidebar, and expandable publication cards. Built from the socialyadept/devportfolio fork and heavily customised for academic use.
Created in 2026
Agentic AI for Autonomous Equity Research
MSc dissertation project at Brunel. A fully functional LLM-powered equity research system using a multi-agent pipeline that retrieves live market data, performs fundamental and technical analysis, synthesises structured reports, and self-evaluates factual consistency. A companion benchmark dataset measures hallucination rates, hierarchical reasoning quality, and factual accuracy against analyst-produced ground truth.
Created in 2025-2026
LLM-Powered SIEM (National Deployment)
First national deployment of an LLM-based SIEM across 200+ federal organisations for Pakistan's NCERT. End-to-end pipeline, Kafka ingestion → LLM-based threat classification (INT8/INT4 quantised inference) → automated alert routing, designed to autonomously process and triage 25M+ logs per day, replacing manual SOC Level 1 analyst operations at national scale.
Created in Feb 2025-Jun 2025