Flagship Projects
Backend & Architecture Focus
I build full-stack applications with a backend-oriented focus: APIs, relational data, application boundaries, deployment, and the decisions behind reliable software.
End-to-End Ownership
Architecture · Implementation · Deployment
Flagship Systems
Oakley Commerce Platform · ObjectID Recognition
Core Focus
Backend · APIs · Relational Data · State Machines
Engineering Standards
Validation · RLS Authorization · Idempotency
About me
Learn. Rethink. Move forward.
Continuous improvement & architectural discipline
Focus
Backend & System Design
APIs, relational data modeling, and reliable distributed architecture
Execution
End-to-End Delivery
Full-lifecycle ownership from requirements to production deployment
Foundation
BSc Computer Science
Strong analytical basis in data structures, algorithms & systems
I’m Hussein, a software engineer with a Computer Science foundation who builds full-stack software systems with a strong focus on backend engineering, APIs, databases, and system design. I work primarily with TypeScript, JavaScript, Node.js, NestJS, Express, React, Next.js, PostgreSQL, and Python, using tools and platforms such as Git, GitHub, Docker, Vercel, Render, and Railway where appropriate.
My work spans the full development lifecycle, from understanding requirements and designing system architecture to implementation, deployment, and post-launch problem solving. I’m particularly interested in what happens beneath the interface: data integrity, API behaviour, application boundaries, performance, failure cases, and the technical decisions that make software reliable in real-world conditions.
I have a BSc in Computer Science and several years of hands-on experience building and delivering software. I’m comfortable working independently when a project requires ownership from idea to deployment, and I work equally well within a team where collaboration, discussion, and shared responsibility matter. I’m adaptable, learn quickly, and approach unfamiliar technologies by understanding how they fit into the wider system rather than simply learning them at surface level.
What I bring is a combination of practical product thinking, engineering discipline, curiosity, and ownership. I care about building software that solves the actual problem, is maintainable after delivery, and continues to work when it leaves the ideal conditions of development.
What I do
I work across the stack to turn product requirements into functional software—connecting user interfaces, backend services, data models, and, where appropriate, AI-powered capabilities.
01
I connect responsive client applications (SPAs, PWAs, Server Components) to centralized backend APIs, authentication pipelines, and transactional data persistence.
Built in practice
Engineered multi-role platforms (Oakley Storefront & Vendor PWA, AWIHF Platform) with shared APIs and centralized business logic.
02
I build server-side architectures around strict RESTful conventions, validation pipelines, authorization policies, state machines, and scheduled background reconciliation.
Built in practice
Designed Express/TypeScript APIs with 3-dimensional state machines, node-cron reconciliation jobs, and automated test harnesses.
03
I design normalized PostgreSQL schemas with strict foreign-key invariants, Row-Level Security (RLS) policies, and transactional integrity guarantees.
Built in practice
Implemented PostgreSQL schemas with database-layer RLS policies and JWT claim helpers in the app schema.
04
I develop applied computer vision workflows across mobile and cloud environments, balancing on-device inference constraints with API-backed fallback pipelines.
Built in practice
Created ObjectID Android app combining quantized MobileNetV3 TFLite inference with Imagga Cloud REST API fallback.
Selected work
These flagship systems make earlier capabilities concrete through real implementation: product architecture, backend decisions, data modeling, deployment constraints, and the trade-offs required to get working software across the line.
01 / Multi-application commerce system
Sole Developer & System Architect
A multi-app commerce platform separating the customer storefront SPA and vendor dashboard PWA, connected to a centralized Node.js/Express API and PostgreSQL with Row-Level Security.
Clear application boundaries, 3-dimensional state machines, database-enforced RLS policies, idempotent draft checkout, and automated end-to-end checkout testing.
Engineering Context
The system was architected to keep customer shopping and vendor administration cleanly separated while centralizing all critical commerce invariants—order state transitions, stock decrementing, payment proof uploads, and scheduled reconciliation—inside a single modular backend.
02 / Android image identification system
Primary Developer
A native Android image identification system combining custom touch bounding-box ROI annotation, quantized on-device MobileNetV3 TensorFlow Lite inference, and Imagga Cloud REST API fallback.
Pragmatic ML architectural pivot balancing on-device mobile hardware constraints with cloud API fallback, paired with robust memory management and defensive error pipelines.
Engineering Context
The project required evaluating the gap between on-device ML compute limits and project delivery timelines. Rather than shipping an unfinished native C++/CMake stack with narrow class coverage, the architecture pivoted to a hybrid approach: cloud-first recognition for high precision with instant on-device TFLite fallback.
Additional Public Work
Public Web Platform
↗Engineered a content-dense platform organizing multiple program domains, field reports, and donation paths, keeping beneficiary information privacy-conscious while maintaining responsive navigation stability.
Public Web Platform
↗Structured a multi-domain organizational platform translating complex grassroots programs, impact metrics, leadership registries, and outreach workflows into a maintainable, high-performance web experience.
Live Production Platform
↗Official live web platform and operational portal for Acholi Women in Health Foundation in Northern Uganda, managing maternal care advocacy, field research data, and active mentorship workflows.
Progressive Web App
↗Built an interactive Progressive Web App featuring dynamic client-side state transitions, user-generated unique proposal links, persistent responses, and audio-visual interaction choreography.
What this proves
Real software engineering isn't just about code that executes on localhost—it is about architectural boundaries, navigating real infrastructure limits, and managing delivery scope.
Oakley separates customer and vendor workflows around a shared backend and PostgreSQL data layer.
The payment flow is presented as a practical workaround, not a fictional direct payment API integration.
ObjectID explains why the delivered version used an external recognition API after local inference proved too broad for the timeline.
How I approach software
I do not start with “which framework should I use?” I begin by understanding the problem, isolating constraints, defining system boundaries, modeling relational state, designing defensive failure modes, and closing the loop with continuous observability.
“What problem are we actually solving, and what is explicitly out of scope?”
Real engineering never begins with framework selection. It starts by establishing operational context, understanding user needs, and defining rigid boundaries around project scope.
“What must the system do, and what limitations dictate the solution?”
Turn ambiguous ideas into concrete engineering specifications across functional rules, non-functional requirements, and real-world environmental limits.
“Which architecture satisfies requirements without introducing unearned operational complexity?”
Architecture is a consequence of requirements and constraints—not fashion. I evaluate alternatives pragmatically, choosing solutions that solve the problem with the least operational overhead.
Selected a single modular Node.js API with strict domain separation rather than microservices, eliminating multi-service deployment overhead while preserving isolation.
“How is data owned, normalized, transitioned, and protected at the database engine level?”
The data model is the architectural foundation of backend systems. I design normalized relational schemas with explicit constraints, foreign keys, and database-enforced security.
“What happens when this fails, times out, or receives malformed/malicious input?”
A senior engineer does not design only for the happy path. I guard boundaries with schema-validated request/response contracts, idempotent mutations, and defensive fallback paths.
Implemented an automatic fallback bridge: runs fast on-device TFLite first; if confidence score < 0.65, escalates to cloud vision, defaulting safely if offline.
“Does the implementation uphold architectural invariants under edge and failure conditions?”
Implementation translates design into readable, modular code with isolated responsibilities, followed by rigorous verification of both normal and degraded behavior.
“How does the system behave under real usage, and what does production feedback teach us?”
Engineering does not end when code compiles or ships. Production telemetry and user interactions reveal real-world assumptions, closing the loop back into problem understanding.
The Engineering Synthesis
Inspect the verified technical record on the Engineering page: multi-dimensional state machines, PostgreSQL Row-Level Security policies, idempotent checkout lifecycles, and live GitHub contribution activity.
Contact
Explore the systems I have built, review my technical work, or get in touch to discuss an engineering opportunity or collaboration.