I build AI-enabled backend systems, automation platforms, product infrastructure, and developer tools. My work combines software architecture, applied AI, data workflows, client integrations, and production operations.
My product experience includes Takaaneh, Paknevis, and Pajoohyar, where I have worked across full-stack development, backend services, browser and desktop integrations, data and AI features, deployment, and operational infrastructure.
- AI systems: LLM applications, AI gateways, agents, RAG, STT, TTS, NLP, and multimodal workflows
- Backend and platform engineering: FastAPI, APIs, WebSocket systems, PostgreSQL, Redis, queues, background jobs, and service integration
- Automation: Telegram and messaging workflows, browser automation, provider routing, retry and failover, and orchestration
- Product engineering: TypeScript, browser extensions, desktop JavaScript applications, dashboards, RTL interfaces, and PDF workflows
- Infrastructure: Linux, Docker, Docker Compose, CI/CD, service operations, logging, deployment, and production troubleshooting
| Project | Engineering focus |
|---|---|
| OCTOPUS | Durable Telegram control plane for AI-assisted software engineering with PostgreSQL, Redis Streams, persistent sessions, browser automation, artifacts, and voice workflows |
| Ajil Unified AI Gateway | Multi-provider AI gateway with routing, failover, observability, WebSocket support, chat, vision, embeddings, image, STT, and TTS capabilities |
| Lemon Editor Extension | TypeScript monorepo combining a Chrome MV3 extension, Express service, rich editing, PDF generation, and RTL/LTR support |
| Speech-to-Text FastAPI Service | FastAPI speech pipeline with FFmpeg preprocessing, local and remote transcription providers, model management, jobs, dashboard, and OpenAPI |
| Answer Sheet OMR | Configurable computer-vision pipeline for answer-sheet alignment, bubble detection, grading, PDF batches, and explainable artifacts |
| EchoLingua | Multilingual TTS and audio-generation workflow with provider routing, recipe-based plans, caching, persistence, and Telegram delivery |
- Triak Trade — guarded AI-assisted signal research and backtesting platform
- Kiwicore — reliable messaging relay with moderation, retries, idempotency, and persistent state
- Gemini API Proxy Service — resilient FastAPI and Cloudflare Worker gateway
- Agentic Election Simulation — reproducible multi-agent simulation research
- Forecasting Oil Price Using AI — market, sentiment, and time-series research pipeline
Languages: Python, TypeScript, JavaScript, SQL
Backend: FastAPI, Laravel, Flask, Express, REST APIs, WebSocket, background jobs, provider integrations
Data: PostgreSQL, MySQL, Redis, SQLite, Pandas, NumPy
AI and automation: LLM applications, AI agents, Hugging Face, PyTorch, TensorFlow, RAG, STT, TTS, NLP, n8n
Infrastructure: Linux, Docker, Docker Compose, CI/CD, systemd, Cloudflare Workers, FFmpeg
Engineering quality: pytest, Ruff, mypy, integration testing, load testing, structured logging, operational documentation
I prefer systems that are:
- observable rather than opaque,
- recoverable rather than memory-only,
- explicit about trust and security boundaries,
- testable at component and integration levels,
- documented for operation as well as development,
- honest about implemented features and current limitations.
I am currently focused on reliable AI workflows, agentic automation, backend architecture, developer tooling, and production infrastructure.