ALIREZA MIRZAEI BACKEND & DATA INFRASTRUCTURE
I’m part of the team turning raw electricity usage data into products that industrial towns and distribution companies actually rely on every day.
A path through backend systems, data engineering, Linux infrastructure, and security.
ELECTRICAL-ENERGY PRODUCTS
I build the backend and data layer behind electrical energy products: ingestion, orchestration, async processing, and the dashboards people rely on. I've also worked on a broader performance refactor and on making background job scheduling more aware of system load. Alongside Baresh and Payesh, I've contributed to a hydropower dam monitoring system and a cryptocurrency miner detection tool.
PLATFORM & DELIVERY
Replaced a manual, ad hoc setup with CI/CD pipelines, a proper ticketing system, containerized services, and more reliable day-to-day operations.
NETWORK SECURITY
Contributed to an internal penetration-testing report tool while practicing Windows and network-security testing.
EDUCATION
Secure Computing Institute.
A few products and experiments that show the kind of problems I enjoy working on.
A utility platform designed around industrial-town customers.
I led Baresh from day one, building the early backend and data layer for a B2C electricity platform serving industrial towns. It handles packages, SMS, notifications, usage dashboards, and rankings, all built on a flexible data layer that can adapt as sources change. Job scheduling adjusts automatically to system load, with a clean data flow between stages so nothing gets lost along the way.
Reporting and demand-management tools for electricity distribution companies.
Payesh runs as its own installation inside each electricity distribution company, so my work has spanned multiple live deployments: features, performance improvements, and infrastructure work tailored to each one. Managers get visibility across feeders, businesses, organizational units, aggregates, and customer classes, with support for usage limits, notifications, and compliance tracking.
Making the model choice an explicit part of delegated agent work.
An extension developed in my Hermes Agent branch: configured capability tags, model categories, defined routing resolution, dependency-aware ordering, validation, tests, and developer documentation.
View Hermes Agent upstreamA ransomware-detection pipeline built for my undergraduate project.
Explored executable-file access patterns and feature engineering with XGBoost and Random Forest. The project reached over 95% accuracy on its dataset.
View repositorySmall public projects for learning through implementation.
Recommerce compares recommender approaches; Breakout explores reinforcement learning in Gym; Manticore is a team data-mining project on employee attrition.
Professional work is described only at a shareable level. Client data, proprietary code, and internal architecture remain confidential.
I’m drawn to work where software has to stay dependable as the data, workflows, and day-to-day operations get more complicated. That means looking beyond endpoints: defining clear behavior, making data flows trustworthy, keeping systems observable, and understanding the infrastructure they run on.
I use AI agents as implementation partners. I bring the architecture, the diagnosis, the constraints, and the review—not just a prompt.
Outside my immediate work, I’m interested in cloud infrastructure, distributed systems, low-level computing, IoT, and Home Assistant.
05 / GET IN TOUCH