Parsa
Harooni
AI software engineer shipping production-grade LLM systems — Python, FastAPI, RabbitMQ, event-driven architectures. Currently building chat & agent infrastructure at fino revOps (Germany) — remote, based in Iran.
Backend engineer focused on putting language models into production where they actually have to work.
My recent focus has been on enhancing chat solutions through the strategic integration of Large Language Models and AI. In the last four months I helped transition a codebase to FastAPI and shipped an event-driven design using RabbitMQ — meaningful gains in throughput, scalability, and cost-per-conversation.
Before that, as a Software Technical Lead at Agentco, I drove AI integration into corporate projects with Python and FastAPI. Across nine years of freelance work and full-time roles across Iran and remote teams, the through-line is the same: ship robust backends, work cleanly across functions, deliver on time.
I care about simple architectures that survive contact with real traffic, observability that doesn't lie, and prompts that fail gracefully when the model has a bad day.
11+ years across full-time, lead, and freelance roles — backend, AI, and team direction.
AI Software Engineer
- Building LLM-driven revenue-ops automation on Python & FastAPI.
- Integrating agentic flows into existing pipelines without breaking SLAs.
AI Software Engineer
- Backend & LLM integration work on a tight delivery cycle.
Team Lead
- Led a team to revamp architecture and transition the codebase to FastAPI for measurable performance gains.
- Implemented event-driven design with RabbitMQ to improve scalability and operational efficiency.
- Designed architecture using LLM models & agents to optimise chat functionality.
Senior Software Engineer
- Led technical direction for corporate projects focused on LLMs and AI solutions.
- Drove innovation with Python, FastAPI, and Django to ship production AI.
- Coordinated across diverse cross-functional teams to deliver high-quality results.
Software Engineer
- Robust, scalable backend solutions in Python, FastAPI, and Django.
- Custom software tailored to client-specific business needs.
- Incorporated LLM & AI capabilities into client projects.
Senior Software Engineer
- Efficient and scalable backend solutions in Python, FastAPI, and Django.
- Cross-functional collaboration to ship high-quality software on schedule.
Technical Consultant
- Built a recommendation system that improved internal business operations.
- Translated unique business needs into technical implementations.
Back End Developer
- Backend infrastructure for a mobile app monitoring children's smartphone usage.
- Secure, real-time data collection with privacy & integrity in mind.
Back End Developer
- Efficient data crawlers handling large-volume information collection & processing.
- Designed scalable systems to ensure data integrity and reliability.
A few representative builds — chat infra, agents, recommenders.
vardast.chat — LLM chat platform
Re-architected the codebase to FastAPI and introduced an event-driven backbone on RabbitMQ. Designed agent & tool-calling layer for a customer-facing assistant.
fino revOps — AI revenue ops
Building LLM-assisted automation for revenue operations workflows; integrating agentic loops into existing data pipelines while keeping latency & cost predictable.
Pixend — recommendation engine
Custom recommender designed around the client's catalog & behavioural signals. Tuned the cold-start path for low-traffic categories.
Kangaroo — child-safety backend
Real-time data collection & processing for a mobile app supervising children's smartphone usage, with strong constraints around privacy & data integrity.
Tooling I reach for — chosen for boring reliability over novelty.
- Python 11y
- FastAPI core
- Django core
- SQL strong
- TypeScript working
- Bash daily
- LLM integration prod
- Agent design prod
- Prompt engineering prod
- RAG & vector search strong
- Evaluation & guardrails strong
- Recommenders past
- RabbitMQ prod
- Event-driven design prod
- PostgreSQL strong
- Redis strong
- Docker daily
- CI/CD comfortable
Largely self-directed — most of what I know shipped to a deadline.
The bulk of my craft has come from nine years of freelance delivery and full-time roles where I owned production systems end-to-end. I keep a steady reading list around distributed systems, applied ML, and language-model behaviour.
If a formal academic record is needed for a contract, I'll happily share it on request via email.