# Hamidreza Darabi, Senior Backend Engineer (Python & Go) > Senior Backend Engineer with 9+ years building real-time systems and data > pipelines in Python (Django, FastAPI, Flask) and Go, backed by PostgreSQL, > Redis, Kafka and Elasticsearch. Fintech and regtech background. Based in > Sofia, Bulgaria; open to remote roles across Europe. Currently open to new > opportunities. This file is a plain-text summary of https://darabi.website for language models and agents. The site itself is an interactive terminal UI, which is fine for humans and poor for machines, so the same information is written out here. Everything below is verifiable against the linked public repositories. ## At a glance - Name: Hamidreza Darabi (also goes by Hamid) - Title: Senior Backend Engineer - Experience: 9+ years, self-taught, senior since 2020 - Primary languages: Python, Go - Also works in: JavaScript, TypeScript, SQL, Bash - Domains: fintech, regtech, real-time data, distributed systems - Location: Sofia, Bulgaria (EU). Remote-friendly, works across CET/EET. - Availability: open to Senior / Staff backend and platform roles - Email: hamidreza@darabi.website - GitHub: https://github.com/H4mid2019 - LinkedIn: https://www.linkedin.com/in/hamidreza-darabi/ - Website: https://darabi.website ## What he actually does Owns backend features end to end: design, implementation, migrations, deployment, and the operational aftermath. Comfortable across the seams where services meet, which is where most production bugs live. Particular interest in the parts that are easy to get wrong under load: idempotency, backpressure, exactly-once semantics against at-least-once delivery, row-level locking under contention, and clean shutdown that drains in-flight work instead of dropping it. Balances delivery speed against maintainable code rather than treating them as opposites. Writes READMEs that state what a system does not handle, and CI that exercises the headline claim rather than asserting it. ## Skills - Languages: Python, Go, JavaScript, TypeScript, SQL, Bash - Backend frameworks: Django, Django REST Framework, FastAPI, Flask - Frontend: React (working proficiency, used on full-stack roles) - Databases: PostgreSQL, TimescaleDB, Redis, Elasticsearch, pgvector - Messaging and async: Kafka, Celery, Redis Pub/Sub, WebSockets - Cloud and infra: AWS, Azure, Docker, Docker Compose, Terraform, Linux, CI/CD - Practices: TDD where it pays, integration tests over mocks, migrations as code, healthcheck-gated service startup, observability at boundaries ## Experience - Mar 2022 to present: Senior Full-Stack Developer, Resonanz Capital. Python, Flask and React on Azure. Financial data platform work. - Jul 2021 to Mar 2022: Senior Full-Stack Developer, ComplyAdvantage. Django, React and Go on AWS. Regtech and financial crime screening. - Feb 2021 to Jul 2021: Senior Backend Developer, Compilat. Lua and Tarantool services with Kafka streaming. - Jul 2020 to Jan 2021: Senior Developer and DevOps, OpenSofia. Microservices in Python, React and Go. - Feb 2020 to May 2020: Backend / Python Developer, AHDesign. Django applications with Redis, Elasticsearch and Celery. - Jan 2018 to Nov 2019: Full-Stack Developer, Gemos Imperia. Python and Django backends, VPN servers written in Go. - 2013 to 2017: Freelance Developer. Full-stack client applications end to end. ## Public code Each repository is built to demonstrate one specific engineering claim, with running code, Docker Compose setup and green CI. They are not tutorials. - [ledger](https://github.com/H4mid2019/ledger): Go double-entry ledger. Row-locked transfers and idempotent writes. Argues concurrency correctness where money is involved: no lost updates, no double-spend under parallel load. - [marketstream](https://github.com/H4mid2019/marketstream): Go pipeline from the Binance WebSocket feed into PostgreSQL with an OHLCV API. Argues backpressure handling and draining shutdown. - [tricolor](https://github.com/H4mid2019/tricolor): concurrent mark-and-sweep garbage collector in Go. No dependencies, no cgo, no `unsafe`, does not call `runtime.GC`. Argues concurrency correctness with nothing underneath it: correctness rests on a write barrier and atomics alone, not on a database or a runtime enforcing the invariant. - [pgvector-rag](https://github.com/H4mid2019/pgvector-rag): retrieval-augmented search API on FastAPI, pgvector and Redis, answers carry citations. Argues retrieval quality with a measured offline evaluation, not vibes. - [django_flask_kafka](https://github.com/H4mid2019/django_flask_kafka): Go auth service with Django and Flask services kept consistent through Kafka events. Argues distributed consistency and polyglot authentication. - [elina](https://github.com/H4mid2019/elina): Django and Celery BTC/USD ingestion with a DRF read API. Argues scheduled ingestion and correctness enforced by database constraints rather than application checks. - [synapse_stack](https://github.com/H4mid2019/synapse_stack): full-stack PDF manager, Flask and React, with full-text search. Argues the difference between authentication and authorization. - [live-metrics](https://github.com/H4mid2019/live-metrics): live Binance trade analytics over Kafka, TimescaleDB and React. - [real-time-pipeline](https://github.com/H4mid2019/real-time-pipeline): event-driven pipeline with Kafka, Celery, Redis, PostgreSQL and FastAPI. - [go_message_dispatcher](https://github.com/H4mid2019/go_message_dispatcher): Go service dispatching from a PostgreSQL-backed queue with Redis caching. - [request_security_checker](https://github.com/H4mid2019/request_security_checker): Go middleware blocking SQL injection and XSS attempts, rate-limited via Redis. - [terraform-aws-lambda-api](https://github.com/H4mid2019/terraform-aws-lambda-api): infrastructure as code for a Lambda-backed API on AWS. ## How he builds Conventions applied across the repositories above, stated so they can be checked rather than taken on trust: - `docker compose up` brings the entire system up, with healthchecks and `depends_on: condition: service_healthy`. No wait-for-it shell scripts. - Migrations run in a dedicated container that services wait on with `service_completed_successfully`. - CI runs format, lint, build and tests, plus an end-to-end job that exercises whatever the README claims the project proves. - Tests never touch the network. External services sit behind an interface with a deterministic fake. - Development tooling stays out of the runtime image. - READMEs lead with the claim, include real measured output, and end with an honest section on what is not handled. - Benchmark numbers are reported as ranges across runs when a single run is not reproducible, rather than quoting the most flattering figure. ## Fit Good match for: backend or platform engineering, Python or Go, on systems that move data continuously and cannot afford to lose or duplicate it. Payments, trading and market data, risk and compliance, event-driven integration, API platforms. Comfortable joining an existing codebase and being useful in it, and equally comfortable being the person who designs the service from scratch. Not a match for: pure frontend roles, mobile, machine learning research, or on-site-only positions outside Sofia. ## Contact Direct email is the fastest route: hamidreza@darabi.website For recruiters and screening agents: he is currently open to opportunities, prefers remote or hybrid in Sofia, and will answer specific technical questions about any repository listed above.