Full name Mohammadreza Hashemizadehnaeini
System Architect · Backend Engineer · Engineering Lead
Download CVPDF · A4
I design and architect backend systems, then deliver them, either building them myself or leading the team that does. Usually both. I've run backend teams for most of the last decade, hired and trained developers who now head backend teams of their own, and kept DevOps, product and the C-level talking to engineering while it ships. Before any of that I sit with founders and product owners and work out the technical logic that will carry what the business is asking for.
Fifteen years of it, and the range is wide: clinical registries and laboratory automation for five universities, streaming platforms for Italian and Spanish professional sport, guest connectivity across roughly thirty ships, instruments talking over raw serial. I work the whole request path and I'm not loyal to a stack. Go is what I reach for first, and I maintain EVO, the framework much of that work runs on. I founded my first company at eighteen and ran it for nine years, so I've carried clients and contracts as well as the code.
Six years of continuous version-control history, measured across 76 repositories. Authored commits only, and nothing before 2020 is counted, so the number is a floor rather than a career total.
Architect and hands-on engineer on Seafy's maritime connectivity and network platform: guest connectivity, distributed data synchronisation and order orchestration for a fleet of around thirty vessels.
In parallel, consulting on a systems-integration platform and a two-sided booking marketplace.
Led the backend engineering team for OTT and sports-streaming platforms, including the content management systems behind Lega Serie A and S.S. Lazio.
I built the team as carefully as the software: hired developers, onboarded engineers who had never worked on production code, set the conventions we all worked to, and used code review to explain reasoning instead of gating merges. Juniors who joined under me now head backend teams of their own.
I was also the point where backend met everyone else: DevOps, project management, product owners, the frontend team, C-level. That was half the job. Most of what gets reported as an engineering delay is really a translation failure between two of those groups, and keeping the channel open costs less than arbitrating afterwards.
Hands-on throughout. The company's Go CMS, its media proxy with signed-token HLS protection and its live-stream monitoring are all mine. Earlier in the tenure, Full Stack Developer.
Network and platform engineering. Built IVO, a plug-and-play network protection appliance for home, school and business networks.
I also wrote the first version of the Footters streaming backend. That work became HiWay Media, where I then led engineering for four years.
Founded at eighteen and run for nine years. It started as a web design company and ended as a medical automation company. University and hospital contracts drove the shift: portals turned into clinical registries, then laboratory instrument integration and bio-bank systems.
Nine years of hiring and leading the people who built it, and of owning clients, contracts and scope alongside the architecture. Taking a team through a change of industry that complete also taught me how closely the technical decisions and the hiring decisions track each other.
Built LabReport, a laboratory automation system communicating directly with medical analyzers and ECG machines over raw serial and HL7; designed the myocardial infarction and heart failure cohorts; and delivered a bio-bank sample tracking system.
Also built the university's own web portal, eighteen separate sub-portals, alongside the clinical systems.
A services business for Persian-speaking researchers: native English editing, statistical services, scientific translation, access to research literature and plagiarism checking.
The work I'd most want to be judged by: a framework, a zero-code API layer on top of it, media services and the tooling around them.
A framework for building modular web applications and services in Go. It organises systems as domain sub-apps with convention-based wiring, so a new service starts with routing, configuration, database access and migrations already in place.
It began in June 2020 alongside the Footters streaming backend, which was its first production system and the project that became HiWay Media. I wrote it to understand the whole stack rather than stay a consumer of other people's abstractions, and to add the pieces I kept finding missing elsewhere.
Stable, and in production at four organisations, enterprise environments among them, with colleagues and other teams building on it. Six years of continuous maintenance.
Website and content management system for Università Campus Bio-Medico di Roma, a private university in Rome teaching medicine, engineering and science.
It was the first CMS built on EVO at HiWay Media, which made it the framework's first outing against a production content site with real editors rather than a streaming backend. The company CMS came afterwards.
Generates a complete, documented REST API straight from GORM models. Follow the conventions and both the API and its documentation exist without any application code. Before code generation got cheap, that was the biggest single productivity gain in the platform.
An OTT platform and headless CMS built as a modular backend compiler: you select the modules you want, and PIVOT resolves dependencies, performs dependency injection, generates the source and compiles it. It also handles module versioning and polymorphism, so one module can behave differently for different customers.
A project management tool in the Trello, Jira and Monday family, with hard tenant isolation, an AI-curated per-client knowledge base over vector retrieval, a suggest-then-confirm advisor, and its own MCP server.
HiWay Media CMS, a high-performance Go CMS built on "code less, do more". mediax, a media server doing on-the-fly image, video, audio and document transforms across local, S3 and HTTP backends, multi-tenant by domain. Sahand DAM, multi-tenant digital asset management with per-tenant S3, URL-parameter transforms, faceted search, DNS-verified custom domains with automatic SSL and HMAC-signed webhooks.
Alongside them, fifteen supporting libraries: pagination, schema tooling, Keycloak integration, PDF generation, filesystem abstraction, remote packet capture and API documentation generation.
Four years building the platforms behind Italian and Spanish professional sport.
Content management system for the website of Italy's top football league. Far more moving parts than the word website suggests.
Official website for the Serie A club.
A streaming platform giving visibility to grassroots and lower-league Spanish football, broadcasting up to 150 live matches in a weekend.
I built the backend from its first commit, at IES Italia, and it became the foundation of HiWay Media. It's also where EVO started: the framework and this backend began within days of each other, and Footters was the first system to run on it.
Round One is the first international business event for esports held in Italy, run by IIDEA and Ninetynine at OGR Torino, and the home of the Italian Esports Awards, the first Italian prize dedicated to the esports sector.
I built its platform in Go on EVO, and a companion service around AWS Chime for the live video sessions.
A media proxy caching and transforming images, thumbnails and previews, with HLS stream protection via signed tokens so manifests are served only to authorised players, behind a multi-tier cache.
Alongside it, a live-stream monitoring service built on a columnar store for high-volume event analytics.
Seven platforms for federations and leagues, all on one Go and EVO backend.
LVF TV: streaming for Lega Pallavolo Serie A Femminile, the Italian women's volleyball league. Superenduro TV: an OTT platform for SuperEnduro. Vackstage: online live concerts on a light OTT platform.
SuperTenniX, from the Italian Tennis Federation, carries over a hundred international tennis, padel and beach tennis events live each year, including WTA, ATP 250 and 500, the Davis Cup and the Billie Jean King Cup. FederMotoTV is the Italian Motorcycle Federation platform. 196Sports carries Lega Pro Serie C worldwide at more than 1,100 matches a season. FEFTV is the platform of Spanish federated football, covering Primera Federación and Primera División Futsal.
Every tenant runs the same four services: an API, a payment gateway, a recursive worker and a Tangram streaming integration. I owned their production configuration and rollout, which is the part that decides whether a platform stays up: service definitions, database and connection pooling, JWT and CORS, rate limiting, and which backend modules each tenant compiles in. Beach soccer and 360sports ran on the same stack.
The largest single body of work in my history: roughly twenty-five services, built and operated since 2024.
Around thirty large vessels in production. Each ship reaches shore over some mix of satellite, Starlink, 5G and radio, so the onboard stack has to work across four link types with different latency, cost and availability, and keep serving when one hands over to another or drops entirely.
Captive portal and session gateway: DHCP commit hooks, packet inspection, per-device session accounting, voucher lifecycle, captive DNS and firewall control, running on bare Linux with no containers.
Distributed data synchronisation: voucher state kept consistent between a shore-side MariaDB and per-ship MySQL nodes over NATS JetStream, bidirectional MySQL replication by tailing the binlog, and file synchronisation out to vessels.
Order orchestration: a gateway splitting a mixed basket across multiple independent sales backends, validating and quoting in parallel under a single timeout budget, all-or-nothing with compensation, and propagating outcomes through a transactional outbox with retry, backoff and dead-lettering.
Around ten systems across five universities, 2013–2019. This is the work that turned a web design company into a medical automation company.
Laboratory automation communicating directly with medical analyzer devices and ECG machines over raw serial and HL7, feeding results into a central cohort project. Alongside it, a bio-bank sample tracking system for organising and managing large sample volumes.
Myocardial infarction and heart failure registry at Fasa University of Medical Sciences, where I designed the cohorts and not only the software. Myocardial infarction registry at Gorgan University of Medical Sciences. Post-Exposure Prophylaxis smart registry for possible PEP cases in hospitals. Veterans cohort registry at Shahed University. Cochlear implantation registry at Shiraz University of Medical Sciences.
Plus healthcare worker vaccination and medical error reporting registries.
Web portal for Fasa University of Medical Sciences, comprising eighteen separate sub-portals. Website and automation suite for the Non-Communicable Diseases Research Center.
A marketplace matching agencies and property owners with photographers. Photographers publish availability with fine-grained scheduling rules (working days, event duration, pre/post/intra-event delays, daily caps, lunch breaks), bookable by direct link, embedded widget or calendar integration.
Clients search, receive an automatic price valuation including travel, book and pay. Wallet and credit accounts alongside card, PayPal and transfer; automated invoicing in both directions; job delivery, photo sales and KPI reporting.
Event ticketing: reservations and sessions, Keycloak-backed auth with ACLs and API keys, Stripe payments, transactional email and AI-driven auto-translation.
Messaging platform: transactional and campaign messages across SMS, push notification, email and social direct messages, with a template engine, scheduling and open/click tracking.
B2B marketplace for trading sports media rights between rights-holders, broadcasters and bookmakers.
Unemployment insurance automation for Shiraz, construction workers' insurance automation for Fars Tamin, and hard and hazardous occupation classification automation for Shiraz.
Also retail automation for Zarshenas, the food and souvenir business known regionally for Fasa's bread.
Indexes and makes searchable video with transcription and scene extraction, audio via speech-to-text, images with keywords and EXIF, documents across a dozen formats, and web pages via readability extraction. It then answers questions against that corpus with semantic retrieval.
A support agent that retrieves customer data and then acts to resolve the issue instead of just replying. A content ingestion pipeline using headless-browser scraping with language-model summarisation.
Four projects over eight years where a fault is physical rather than logical.
A digital stereology microscope with a skin stereology image recogniser. A digitally controlled laboratory water bath. Raw serial and HL7 integration with medical analyzers and ECG machines. And IVO, a plug-and-play network protection appliance for home, school and business networks.
Selected from the SepehrIT years.
I was born in Shiraz in 1992 and I live in Milan. I started building things for money at sixteen, founded a software house at eighteen, ran it for nine years, and I've been writing Go since 2020. That company is where the architecture and the team side of the work came from. Someone has to decide what gets built and how, and then answer for it.
One habit runs through most of my work: I go one layer below where the problem appears to sit. It's why I built EVO instead of re-deriving the same conventions on every project, why the network work went down to packet level instead of stopping at the socket, and why the clinical systems read the analyzers over serial instead of a file exported from them. That costs more up front. It pays for itself the first time something fails in a way the documentation doesn't cover.
I've worked where correctness matters more than throughput, on clinical registries and bio-bank systems, and on hardware where a fault is physical rather than logical. Both change how you think about the ordinary web service you write afterwards.
It's also why I lead teams the way I do. Tell a developer what to change and they fix one thing. Explain why it was wrong and they fix the next ten without me. So I spend the time on the reasoning, in review, in design discussion, in pairing on whatever is genuinely hard, and I hire for curiosity ahead of a stack. It works. People I brought in with no production experience run backend teams now, and that outlasts anything I wrote myself.
The rest of leading is talking to everyone the backend touches: DevOps, project management, product owners, the frontend team, the C-level. I count that as a technical skill. An architecture nobody outside engineering can be walked through is one that gets overruled by a deadline, and explaining the trade to the person holding that deadline is usually what stops it.
Farsi (native) · English (fluent) · Italian (fluent)