Dublin-based · Open to backend roles
I’m Hussein, a backend engineer building reliable Java and SQL systems - with AI and analytics depth for data-heavy problems. The goal is boring production days.
Selected projects · Backend, data, AI
Five projects, each built around a real constraint: real-time voice AI, local-first systems, production Java, applied AI, and cloud infrastructure.
Co-building an early-stage real-time voice AI platform. I own the initial engineering foundation across LiveKit conversational pipelines, RAG with PostgreSQL and pgvector, agent tool execution, the Next.js operations interface, and Supabase persistence.
Local-first macOS activity intelligence: a native menu-bar collector writes privacy-filtered foreground, document, Safari, and idle events to append-only JSONL; an Obsidian plugin turns them into focus blocks, context-switch analysis, goals, comparisons, and weekly reviews. The collector stays under 20 MB of active memory with negligible battery impact. GPT-5.6 assisted the implementation; I owned the architecture, engineering decisions, and validation.
Spring Boot REST API for redirect resolution, custom aliases, and click analytics, backed by PostgreSQL with connection pooling. Even short links deserve reliable infrastructure.
Django education platform with a real-time face-recognition attendance pipeline: 95% accuracy and 80% less marking time. Attendance no longer gets the first ten minutes of class.
Cloud deployment backend for OCI with REST infrastructure provisioning, real-time WebSocket build logs, and a supporting CLI. Deployment logs arrive before the demo ends.
Oracle Financial Services Software · 2023–2025
At Oracle Financial Services Software, I worked on core banking systems where a slow batch job or silent data error had immediate downstream consequences. Banks are not famous for enjoying surprises.
Rebuilt a critical reporting path with parallel SQL execution and better source selection, eliminating business-of-day delays.
Architected 64-way partition-wise processing across eight source tables for the highest-volume jobs—145M records processed.
Resolved more than 80 recurring failure patterns across core banking modules through targeted Java fixes and defensive refactors.
Automated SQL and Bash recovery flows to detect, retry, and log file-processing failures that previously needed manual intervention.
Led a six-engineer team building a production web platform, owning architecture, code reviews, delivery milestones, and stakeholder coordination from first commit to deployment.
Capabilities · Proven in production
Backend engineering is the core. Data platforms and applied AI extend what I can build when the system needs more than an API. The stack is a toolbox, not a personality.
Production Java services, failure recovery, high-volume batch processing, and APIs designed around real operational constraints.
Query tuning, partitioning, warehousing, and analytics pipelines built to keep large workloads understandable and fast.
Applied AI systems spanning real-time voice, retrieval infrastructure, computer vision, and the practical infrastructure needed to deploy and operate them.
Backend first · Broader by design
I think in data flow, query plans, partition keys, and failure modes before framework choices. Java is my primary tool; SQL and Python cover the rest of the path from durable APIs to analytics and applied AI. If the data takes the scenic route, I want the query plan.
UCD Michael Smurfit Graduate Business School · Expected 2026
KJ Somaiya College of Engineering · CGPA 9.69/10
Co-authored a low-cost classroom attendance system validated across varied photo conditions.
Read paper ↗