Home / Portfolio

Portfolio

Systems built to enterprise standards, now built for businesses like yours.

Every project below solved a real operational problem — too many spreadsheets, too much manual matching, too little visibility into what's actually happening in the business.

Real-time ERP · 2024–Present

Nexus Enterprise Platform

The problem: Different teams working from different versions of the truth — one dashboard says one thing, another says something else, and nobody fully trusts the numbers.

What was built: A real-time ERP system architected around a single source of truth across every business domain, with role-based access control, full audit logging, and transactional integrity — so an update in one place shows up everywhere, instantly, for every concurrent user.

Result: one live system that every department trusts, instead of five spreadsheets that disagree.

Real-time syncRBACAudit loggingSupabaseLogistics workflows
Multi-Tenant SaaS · 2024–Present

Vehicle Lifecycle Management (VLM) Platform

The problem: A vehicle-service business needed software that worked in the field — with patchy internet — and could bill customers automatically instead of manually tracking service contracts.

What was built: A multi-tenant SaaS platform with row-level security for data isolation, an offline-first sync engine with conflict resolution, a rule-based workflow engine for AMC billing and service lifecycles, and a hybrid cloud storage setup for documents and media.

Result: field teams keep working with no signal, and billing runs itself.

Multi-tenant (RLS)Offline-firstAutomated billingCloudinary + Cloudflare R2
Data Automation · Enterprise Consulting

Financial Data Pipeline & Entity Resolution

The problem: A financial enterprise was manually reviewing large volumes of transaction data to find matching records and unusual activity — slow, error-prone, and impossible to scale.

What was built: An end-to-end automated pipeline — ingestion, transformation, and storage — with entity resolution using fuzzy matching and similarity scoring, distributed processing for large datasets, ML-based anomaly detection, and centralized monitoring for full visibility into the pipeline.

Result: matching and anomaly checks that used to take days now run automatically, with monitoring built in.

Apache NiFiSpark / PySparkFuzzy matchingAnomaly detectionSplunk monitoring
Cloud Migration · Retail

Retail Data Processing & Cloud Migration

The problem: A retail and inventory data setup had outgrown its original platform and needed to move to more scalable, analytics-ready infrastructure without losing data integrity along the way.

What was built: A migrated and restructured pipeline moving retail and inventory data onto cloud infrastructure, with schema validation, transformation, and analytics-ready outputs, plus monitoring for ongoing observability.

Result: retail and inventory data ready for analytics and future ML use, with no gaps in the move.

AWS S3 + SageMakerSchema validationSplunk
Client Work · Ongoing

Business Websites, Portals & CRMs

The problem: Small and growing businesses needed a professional online presence and simple internal tools — without enterprise budgets or timelines.

What was built: End-to-end digital solutions including business websites, customer portals, and internal management systems — with lead tracking, CRM functionality, and service-operation workflows tailored to how each business actually runs, plus SEO and analytics integration to grow visibility.

Result: a professional online presence and simpler day-to-day operations, sized to fit the business.

Custom websitesCRM & lead trackingSEO & analytics
Enterprise Backend · 2019–2022

Smart Operator — Distributed Backend Systems

The problem: A growing enterprise platform needed backend systems that could scale, separate read and write workloads cleanly, and migrate safely from a monolith toward microservices.

What was built: Backend systems using CQRS and MVC patterns for scalability, optimized data access reducing query times from minutes to seconds, multi-tenant configuration via MongoDB, and a monolith-to-microservices transition on a distributed, fault-tolerant architecture — including ML-based fuel consumption models for predictive analytics on vessel operations.

Result: a backend that scaled with the business instead of slowing it down.

.NET CoreCQRSMicroservicesML predictive models

Have a similar problem?

Whether it's five spreadsheets that disagree or a product that needs to exist, let's talk about what it would take to build it.

Start a Project