At Hisani, I’m architecting and building a TypeScript and React SaaS platform for the horse-racing ecosystem, spanning Stopwatch, Planner, Marketplace, Hire, Stable, and Syndicate. I own its architecture, implementation, AWS deployment, product decisions, and development guided by user feedback.
At The British University in Dubai, I conducted federated-learning experiments using FEDn and analyzed the results. I also co-authored research submitted to Information Sciences on imbalanced data streams under label delay and concept drift, and I’m building SCII, a research prototype for context-aware citation impact.
During Google Summer of Code 2025 with the Eclipse Foundation, I orchestrated a modular multi-agent system with repository context, retrieval-augmented generation, and an MCP server. I also designed and implemented the REPD algorithm for software defect prediction and anomaly detection, and built data pipelines and GitHub Actions workflows for reproducible training and automated analysis.
I built RacingStatto, a live horse-racing analytics platform, and Agentify, a CLI for creating repository-aware coding agents. For MathVizAI, I engineered a system that turns math problems into educational videos using LLMs, Manim, RAG, and a self-correcting multi-agent workflow.
