Backend & APIs
Designing business logic, web services, and REST APIs with Django, FastAPI, Flask, and Laravel.
01
About me
Backend developer focused on building and maintaining production services, REST APIs, and data-driven web applications with Python and PHP. I work with FastAPI, Django, Flask, and Laravel, implement Agentic AI workflows with OpenAI, LangChain, and LangGraph — including human-in-the-loop controls and tool calling — and run services with Docker and Linux. WordPress/WooCommerce plugin development and integrations sit alongside the core backend work as complementary experience.
Data & Storage
Working with PostgreSQL, MySQL/MariaDB, MongoDB, SQL Server, and Redis; file handling with Amazon S3 and object storage.
Infrastructure & Deployment
Hands-on work with Docker and Linux, runtime preparation, delivery-cycle support, and VoIP infrastructure experience with Asterisk, Issabel, and SIP.
WordPress & WooCommerce
Designing sites and developing WordPress/WooCommerce plugins; customizing user experience and integrating with internal services.
02
Work experience
Backend Developer · DevOps
Working on Bijak products across Python backend services and DevOps responsibilities, with additional work on Agentic AI systems and the company website and WordPress/WooCommerce plugins.
- Developing and maintaining product backend services and APIs with Python, FastAPI, and Django in production.
- Handling DevOps responsibilities across Docker, Linux, and Git, including containerized runtime environments, deployment-related configuration, and service operation and troubleshooting.
- Contributing to Agentic AI systems and LLM-based workflows, including model orchestration, tool calling, external API integration, and human-controlled workflow stages.
- Designing and developing the company website and WordPress/WooCommerce plugins integrated with internal and external services.
Web Application Developer
Built data-driven systems including a correspondence workflow and an organizational phonebook, using Flask and Python with MongoDB, MySQL, and SQL Server for high-volume data.
- Implemented web services and features with Flask and Python for data-driven systems, from correspondence tracking to large contact-dataset management.
- Worked with MongoDB, MySQL, and SQL Server for storage, search, and data management in high-volume scenarios.
Web Application Developer
Pol Darmanegaran Iranian Kohan
Contributed to Laravel-based web services and product features.
- Implemented service logic and user-facing features with Laravel, PHP, JavaScript, and MySQL.
- Contributed to interface implementation with Bootstrap 5 and PWA-oriented parts of the project.
Earlier Experience — Web & WordPress
WordPress Web Designer
Designed and implemented an organizational website for a legal group with WordPress, HTML5, and CSS3.
Web Designer & SEO
Designed a data-recovery services website and optimized page structure and content for SEO.
Web Designer
Designed and implemented a website with WordPress, HTML, and CSS, aligning presentation with content needs.
03
Agentic AI, applied projects & open source
Selected projects
Agentic AI project2026
AI Agent for Blog Content Production
Multi-step AI agent for blog content production built with Python, OpenAI, LangChain, and LangGraph. The workflow is stateful, letting a human approve or edit outputs at review gates, add instructions, and resume execution from the same state.
User InputLangGraph AgentLLM (OpenAI)Tool / StageHuman ReviewResult
- Orchestrated independent content-production stages and managed workflow state with LangGraph.
- Implemented human-in-the-loop controls for pausing, reviewing, revising, and resuming the process in a controlled way.
Engineering case study
Problem
Blog content production spans several dependent stages, and each stage output needs review before continuing; without human oversight, an early-stage error would invalidate the entire chain.
Architecture / Approach
Each stage was implemented as an independent node in a LangGraph graph, with full workflow state persisted. At review gates execution pauses and resumes from the same state after the user approves or edits.
Engineering decisions
LangGraph was chosen for orchestration and state management to keep resumable execution and stage branching simple; LangChain abstracts LLM invocation.
Stack
Python · OpenAI API · LangChain · LangGraph
Agentic AI project2025
AI Agent for Address Validation & Geolocation
Intelligent address-processing system for order registration that validates city and province existence, verifies city-province membership, and normalizes address structure. After the LLM evaluates the address text, the agent uses tool calling to send it to the Neshan API, retrieve candidate coordinates, reverse-geocode them, and compare the returned address with the input. Precise coordinates are stored only above a similarity threshold; otherwise city reference coordinates are used.
Address InputLLM EvaluationTool: Neshan SearchReverse GeocodeSimilarity CheckStore / Fallback
- Designed a validation and normalization pipeline for incomplete or inconsistent city, province, and address data.
- Integrated LLM tool calling with the Neshan API and implemented similarity-threshold decisions with a reliable fallback.
Engineering case study
Problem
At order time, addresses are often incomplete or inconsistent, and incorrect geocoding can disrupt logistics.
Architecture / Approach
The agent first evaluates the address text with an LLM, then retrieves candidate coordinates via tool calling to the Neshan API and verifies them with reverse geocoding.
Engineering decisions
The final decision uses a similarity threshold between the input address and the reverse-geocoded result; below threshold the city reference coordinates are stored as a fallback to avoid persisting incorrect coordinates.
Stack
Python · LLM · Tool Calling · Neshan API · Geocoding
Applied project2024
Correspondence Workflow System
Internal system for registering, sending, and tracking correspondence with direct scanner integration, attachment handling, receipt logging, forwarding, and replies. Built with Flask, MongoDB, Amazon S3, and OCR and evaluated in practical testing for at least five million records.
- Covers the full internal correspondence flow from registration and scanning through receipt, forwarding, and reply.
- Manages attachments and object storage alongside the system's operational data.
Applied project2024
Organizational Phonebook Application
Web application for managing a large contact dataset with custom internal fields, pagination, and advanced search for fast retrieval, implemented with Flask and MongoDB and suited to scenarios of around ten million records.
- Managed custom internal contact fields alongside paginated browsing.
- Implemented advanced search for rapid retrieval within the contact dataset.
- Designed for high-volume datasets and scenarios approaching ten million records.
Other projects
Pol Drug Store
Web services for pharmacy-related listings and pharmaceutical consultation bookings, built with Laravel, PHP, and MySQL.
Open-source projects
All projects on GitHubbijak-woo
WordPress/WooCommerce plugin and utilities for the Bijak ecosystem.
PHP · JavaScript · CSSham-hesabMulti-technology web project with a Python backend, TypeScript/JavaScript frontend, and Docker-based development setup.
Python · TypeScript · Dockerjalali-calendarPython utility for working with Jalali calendar dates.
PythoncropXWeb utility for image and TIFF-file operations.
Python · Shell · DockerRadiusAdminPanelFlask-based web administration panel for RADIUS.
Python · Flask · JavaScriptPersianScannedDocOrientationCorrectionOCR-assisted Python utility for detecting and correcting the orientation of Persian scanned documents.
Python · OCR · Jupytermultipage-tiff-compressorPython utility for compressing and converting multi-page TIFF images.
Python · TIFF