Hey, I'm Uriel.
I focus on concurrency, parallelism, and software architecture, and on combining that foundation with machine learning: reading datasets, choosing what to apply, and interpreting what the outputs actually mean.
About
I'm a backend developer focused on Go, building systems that need to behave correctly under real load, not just work on the happy path. That means deliberate data modeling, correct concurrency guarantees (I know when to reach for a mutex, a channel, or neither, leaving the guarantee to the database), and architecture decisions that are documented, not just implied by the code. I'm bringing that foundation together with statistics and applied data science, aiming to make architecture decisions driven by data, not just intuition.
EU citizen (Italian), open to remote roles or relocation to NL/DE.
Experience
Jan 2026, Present
Promoted from intern to full-time. Delivering full-stack features across 5+ client projects (Python/Django, TypeScript/React, NestJS, PostgreSQL, AWS), including a CRM-style engagement system with an automated, behavior-triggered email pipeline, and contributing to a UNICEF and Médecins Sans Frontières digital health initiative.
Aug 2024, Jan 2026
Delivered full-stack features (React, Django) on custom e-commerce platforms, with a focus on performance and scalability.
Jul 2026, Present
Rebuilding a client’s e-commerce platform end to end. Go backend using goroutines and channels for concurrent order processing, Next.js frontend built with SEO as a priority.
May 2024, Jul 2024
Built a Flask API with an asynchronous Celery pipeline to process video submissions through a facial recognition and analysis flow, delivered to and adopted by the medical team overseeing the project.
Technologies I work with
Work
Ticket sales backend in Go, built to solve 3 real problems in high-load systems: N:N modeling, real idempotency enforced via a database constraint instead of an application-level check, and concurrency, where a simulated race condition was first solved with a mutex, then with a worker pool over a channel, with a benchmark comparing both approaches. Load tested with 2,000 concurrent requests, with observability through structured logs.
Full-stack production tracking system for the textile industry, letting companies manage outsourced manufacturing batches across multiple production stages. Go backend with real-time status updates over WebSocket, plus a QR-code flow so outsourced facilities can confirm stage completion without manual reporting. Built solo, backend to frontend.
Freelance full-stack engagement rebuilding a client’s e-commerce platform from scratch, owning backend and frontend architecture. Go backend using goroutines and channels for concurrent order processing, Next.js frontend built with SEO as a priority. Client project, private repository.
Remaining useful life prediction system for aircraft engines using the public NASA turbofan dataset, combining a Go and WebSocket API to serve real-time predictions with a predictive model in Python. Focused on MLOps: model comparison, versioning, and monitoring.
What I do
I choose between a mutex, a channel, or neither, leaving the guarantee to the database, and measure the real cost of each decision instead of defaulting to one.
Ports and adapters, schema decisions explained (why NUMERIC, why no ORM, why cascade here and not there), code that explains itself.
APIs tested under real load: a load test comparing 2,000 concurrent requests, not assumptions.
In progress: applying predictive models to real sensor data (NASA turbofan dataset) to estimate the remaining useful life of equipment.
Let's build something great.
Open to new roles and interesting projects.