Skip to content
SS·
Open to work4+ years · Backend & Architecture

Shayan Sadeghi

شایان صادقی

Backend Developer & Solution Architect

I design and build scalable backend systems — microservices, asynchronous pipelines, and high-availability APIs that stay fast under real production load.

Four years across telecom and fintech, with a growing focus on AI-assisted backend engineering.

View ProjectsDownload Résumé

Tehran, Iran

4+
Years of experience
Backend engineering across telecom and fintech
45,000
Representatives served
High-load service-center microservices at MTN Irancell
92%
Faster API responses
Async, task-based rewrite cut ~39s calls to ~3s
90+
Learners mentored
Three Python/Django cohorts at Maktab Sharif

01 — About

Systems thinking, from mechanical engineering to distributed software

I'm a backend software engineer based in Tehran with four years of experience turning complex business problems into reliable software. Today I work at MTN Irancell — one of Iran's largest telecom operators — leading backend work on a nationwide platform spanning 500 service centers and maintaining high-load microservices used by 45,000 registered representatives.

My work centers on what happens behind the endpoint: asynchronous processing, message-driven design with RabbitMQ and Celery, careful database modeling and query performance, and production observability with Elasticsearch, Prometheus, and Grafana. I also lean on AI agents to speed up analysis, debugging, and testing — while keeping engineering review and ownership on my side. I care about clean architecture: code the next engineer can read, extend, and trust.

I started in mechanical engineering, where my thesis applied machine learning to fault detection in wind turbines — and that systems mindset never left. I'm now pursuing an M.Sc. in Computer Engineering (AI), and I've mentored 90+ learners through Python and Django cohorts. Teaching is how I sharpen my own understanding.

02 — Experience

Where I've built things

Four years of backend engineering — from an AI trading startup to one of Iran's largest telecom operators — plus teaching and mentoring along the way.

Backend Software Engineer

MTN Irancell

Sep 2023 — Present · Full-time · On-site

A leading telecommunications provider offering mobile, data, and broadband services across Iran.

  • Visitor Management Platform (team lead): led backend development for a nationwide platform spanning 500 service centers, covering Salesforce admin workflows, mobile APIs, and RBAC. Onboarded 2 engineers and refactored legacy queries with targeted indexing, cutting visitor search report load time by 70%.
  • SIM Management Platform: designed and implemented a caching layer for user role and permission data pulled from Irancell's external core services — eliminating redundant calls on every login and authorization check and reducing average lookup latency by ~65%.
  • DPOS & account microservices: maintain and optimize high-load services supporting 45,000 registered service-center representatives nationwide (~12,000 daily active users), improving performance and reliability through database query optimization and clean, maintainable code — cutting failed transaction rates by 30%.
  • Use Elasticsearch for production issue investigation, bringing mean time to resolution down from 2 hours to under 20 minutes.
PythonMicroservicesREST APIsRedisPostgreSQLElasticsearchRBAC

Backend Developer

Mahda

Mar 2022 — Sep 2023 · Full-time · On-site

An AI startup building advanced technology for stock-portfolio management and trading automation.

  • Built an options-market Telegram bot on Celery and Redis, replacing repeated client-side polling with scheduled asynchronous processing.
  • Migrated synchronous workflows to an asynchronous, task-based architecture, reducing average API response times from ~39s to ~3s (roughly 92%) and making the endpoints far more efficient to work with.
  • Applied design patterns to keep the new asynchronous flow maintainable.
  • Built Prometheus and Grafana dashboards covering service health, background-task execution, and application performance.
PythonCeleryRedisasyncioPrometheusGrafana

Teaching & Mentoring

Technical Mentor & Python Instructor

Maktab Sharif & Tehran Technical Complex

May 2022 — May 2024 · Part-time · Remote / Tehran

One of Iran's best-known software engineering bootcamps, alongside a long-established technical training institute.

  • Mentored 90+ learners across three Python/Django cohorts through code review, debugging, and practical project guidance.
  • Taught Python and backend engineering fundamentals — algorithm design, object-oriented programming, and API design.
PythonDjangoCode reviewAPI designOOP

03 — Architecture

How I think about backend systems

Good architecture is mostly about what happens under load, at 3 a.m., and six months after launch. These are the areas I keep coming back to.

API layerservicesrabbitmq / celeryworkerspostgresqlredisprometheus → grafana

Scalable Backend Systems

Stateless services and load-aware design that keep latency predictable as traffic grows from thousands to millions of requests.

Microservices

Decomposing platforms into focused, independently deployable services with clear contracts and well-defined ownership.

Async & Message-Driven Systems

Celery, RabbitMQ, and asyncio pipelines that move heavy work off the request path and keep systems responsive under load.

High-Availability APIs

REST and gRPC interfaces engineered for uptime — versioned, documented with OpenAPI, resilient to downstream failure.

Performance Optimization

Profiling, Redis caching, and index-aware query tuning with measured outcomes — like a 70% faster report and a 65% cut in lookup latency.

Observability & Debugging

Prometheus metrics, Grafana dashboards, and Elasticsearch log search that turn a two-hour production investigation into a twenty-minute one.

Authentication & Authorization

Custom authentication flows and role-based access control (RBAC) that keep APIs secure without slowing teams down.

Database Design & Caching

PostgreSQL schema design, MongoDB where it fits, and Redis caching layers shaped around real access patterns.

04 — Projects

Selected work from my GitHub

Chosen for their architecture and code — not their star counts. Each one has a short case study with the technical decisions behind it.

claude-telegram-bot

2026

Personal project · Public repository

A self-hosted Telegram bot powered by the Claude Agent SDK — text, voice, photos, and video in; analyzed media and generated files out.

TypeScriptNode.jsClaude Agent SDKMCPTelegram Bot API
Read case studyView on GitHub

crypto_service

2022

Personal project · Public repository

A compact real-time price-watch service: WebSockets in, dynamically scheduled Celery tasks in the middle, Redis-backed push notifications out.

PythonDjangoDjango ChannelsCeleryCelery Beat
Read case studyView on GitHub

Ecommerce_django

2023

Earlier learning project · Public repository

A full-surface Django e-commerce platform from my learning years — auth with OTP verification, DRF APIs with custom permissions, orders and discounts, tests, and one-command Docker setup.

PythonDjangoDjango REST FrameworkPostgreSQLDocker Compose
Read case studyView on GitHub

05 — AI & Research

Where backend engineering meets machine learning

My path into AI runs through engineering: a mechanical-engineering thesis on fault detection, a master's degree in artificial intelligence, and hands-on AI-agent side projects.

Academic research · B.Sc. thesis

Machine-Learning-Based Fault Detection in Wind-Turbine Engine Disks

An academic machine-learning study on classifying faults in wind-turbine engine disks and evaluating data-driven approaches to predictive maintenance.

  • Preprocessed high-dimensional sensor data using SVD and POD dimensionality-reduction techniques.
  • Developed and compared K-means and MeanShift clustering against a supervised MLP neural network for fault classification.
  • Fine-tuned the MLP to significantly outperform the unsupervised models, informing predictive-maintenance strategies.
  • Investigated sensor technologies for continuous monitoring in harsh environments — focusing on speed, accuracy, and efficiency for the motor and aviation industries.

Methodology at a glance

  1. Sensor data
  2. SVD · POD
  3. K-means · MeanShift · MLP
  4. Fault classification
  5. Maintenance insight

M.Sc. Computer Engineering — Artificial Intelligence

In progress at Islamic Azad University (expected 2027) — deepening the theoretical side of machine learning alongside production backend work.

Applied AI-agent engineering

Beyond academia, I build with AI in practice — most recently a self-hosted Telegram bot powered by the Claude Agent SDK that handles text, voice, and media end to end.

See the project
Mechanical engineering taught me to model physical systems; backend engineering taught me to build digital ones that don't fall over. Machine learning is where the two meet — data-driven decisions running on infrastructure that scales.

06 — Skills

Tools I work with

Grouped by how they're actually used — no percentages, no progress bars. Depth comes from four years of production work.

Languages

Python has been my daily driver for four years, with SQL right behind it.

PythonSQLJavaScriptHTML / CSS

Backend Frameworks

Django for product platforms, FastAPI for async services.

DjangoDjango REST FrameworkFastAPIFlask

Databases & Storage

Schema design, index-aware query tuning, and caching strategies.

PostgreSQLRedisMongoDB

Messaging & Realtime

The backbone of the async, event-driven systems I build.

RabbitMQCelerygRPCWebSocketsasyncio

API Engineering

RESTful APIsOpenAPI / SwaggerAuthentication & Authorization

Architecture & Practices

MicroservicesClean ArchitectureDesign PatternsUnit TestingAgile / Scrum

DevOps & Infrastructure

DockerLinuxNginxGit

Observability & Search

Dashboards I've designed are watched daily in production.

ElasticsearchPrometheusGrafanaApplication loggingService-health monitoring

AI & Machine Learning

From thesis research and graduate study, plus AI agents used daily to speed up analysis, debugging, and testing.

AI Agents (Claude Agent SDK)Classification & ClusteringNeural Networks (MLP)Dimensionality Reduction (SVD · POD)

Familiar with

Working exposure rather than daily production ownership.

MinIOKubernetesCI/CD pipelines

07 — Education

The academic path behind the engineering

An unusual route — mechanical engineering, then software, now artificial intelligence — that turned out to be the point: systems thinking applies everywhere.

In progress

M.Sc. Computer Engineering — Artificial Intelligence

Islamic Azad University, Electronic Branch

2024 — expected 2027

Graduate study focused on machine learning and intelligent systems, pursued alongside full-time backend engineering.

Completed

B.Sc. Mechanical Engineering

Islamic Azad University, Science and Research Branch

2018 — 2023

Thesis: machine-learning-based fault detection in wind-turbine engine disks — the project that pulled me toward data-driven software.

08 — Writing

Blog

Notes on backend engineering, architecture, and applied AI.

Articles coming soon

I'm preparing in-depth, practical posts — the kind I wish I'd found when debugging queues at 2 a.m. No fluff, no reposts.

Topics on the way

Backend EngineeringSystem ArchitecturePythonDjangoFastAPIPostgreSQLRedisRabbitMQAsync SystemsObservabilityAI & ML

09 — Contact

Let's talk

Whether it's a role, a project, or a technical question — my inbox is open. I usually reply within a day or two.