Professional Summary
Full-Stack Software Engineer and former Tech Lead with 13 years of experience turning complex problems into reliable, production-grade software across frontend, backend, automation, and AI. Works end-to-end with React, TypeScript, Next.js, Python/FastAPI, Java/Spring Boot, PostgreSQL, Redis, and RabbitMQ. Most recently led development at Neshan, one of Iran's largest mapping platforms (estimated 4-6M monthly visits), cutting backend response times by ~60% with Redis caching, building an automated RabbitMQ content pipeline across ~4M pages, and improving Core Web Vitals by 40%. Currently pursuing an M.Sc. at TU Clausthal in Germany while building independent products including LeitnerPlus, Speech with Me, and SeeMirror.
Personal Projects
LeitnerPlus
Next.js, React, TypeScript, Tailwind CSS, Supabase, PostgreSQL, Row Level Security, Supabase Auth, Supabase Storage, FSRS, Vercel
Developed a production-oriented full-stack spaced repetition platform with Supabase-backed authentication, PostgreSQL data modeling, RLS-secured user data, private image storage, signed URL access, and server-side FSRS review scheduling. Implemented card creation/editing/deletion, learning spaces, tags, categories, review logs, dashboard analytics, SEO landing pages, legal consent flow, Google OAuth, light/dark mode, and Vercel production readiness.
Speech with Me - AI Speaking Coach
Electron, FastAPI, SQLite, MLX-Whisper, Apple Speech, Local AI Providers, OpenAI-compatible APIs
Privacy-focused AI language coaching desktop app with local and hybrid providers including MLX, LM Studio, Ollama, OpenAI-compatible APIs, and Gemini. Built local speech recognition, native Apple Speech bridging, TTS, Hugging Face model management, setup diagnostics, SQLite sessions, chat history, learning items, summaries, and multi-language workspaces.
PyPilot
Electron, React, TypeScript, Vite, electron-store, Zod, Vitest, electron-builder
Cross-platform Python script launcher supporting Python, Conda, Poetry, uv, Hatch, and custom runtimes. Implemented runtime detection, command preview, project-root inference, environment injection, built-in console execution, system-terminal/background run modes, process-tree termination, tray integration, launch-at-login, onboarding, tests, and packaged macOS/Windows/Linux releases.
Media to Subtitle
Python, CustomTkinter, MLX-Whisper, OpenAI Whisper Fallback, FFmpeg, PyInstaller
Local transcription tool converting audio/video files into TXT and SRT without cloud dependency. Implemented batch processing, model selection, FFmpeg probing/enhancement, noise-reduction profiles, subtitle timestamp generation, export flows, Metal cache cleanup, update checks, keep-awake support, and optional local LLM refinement.
S-Player
Electron, JavaScript, Subtitle Parsing, Local Playback, Media Player
Subtitle-heavy media player for language learners and researchers. Implemented playback resume, playlists, repeat modes, keyboard seeking, subtitle-synced transcript navigation, text selection/copying, sleep prevention, and robust parsing/merging for SRT, VTT, ASS/SSA, SBV, TTML/DFXP/XML, SAMI/SMI, and LRC subtitle formats.
Clipboard Reader
Swift, SwiftUI, AppKit, AVFoundation, WidgetKit, AppIntents, Carbon, CoreGraphics
Native macOS accessibility utility that reads clipboard text aloud with AVSpeechSynthesizer, language detection, multilingual voice profiles, automatic voice resolution, global shortcuts, menu-bar UI, launch-at-login, settings import/export, App Group shared storage, WidgetKit status widget, and update checking.
NeoStream Video Grabber
JavaScript, Electron, React, Chrome Manifest V3, Zustand, FFmpeg, HLS/M3U8
Desktop download manager plus Chrome extension for detecting and downloading media from web pages. Built with a JavaScript monorepo, Electron, React, Chrome Manifest V3, Zustand, and FFmpeg. Implemented DOM/network media detection, retry queues, token-protected localhost communication, direct downloads, authorized HLS/M3U8 downloads via FFmpeg, progress parsing, pause/resume/cancel/retry, persistent history, duplicate prevention, tray behavior, sleep prevention, and cross-platform packaging scripts.
SeeMirror - Real-time EEG Neurofeedback
Python, DSP, ADS1299, SPI/I2C, Real-time Signal Processing, Spectral Analysis, Noise Reduction, Embedded Systems, Linux
A neurofeedback platform (seemirror.ai) built around a single idea - become the observer of your own mind - capturing live brain activity and mirroring it back to the user as real-time, on-screen feedback. Engineered the full signal chain end-to-end: custom ADS1299-based acquisition hardware, a Python DSP pipeline, and high-fidelity, low-latency capture over SPI/I2C with real-time filtering, noise reduction, and spectral analysis plus robust streaming for stable continuous recording. Owned hardware-firmware integration and used AI-assisted diagnostics to speed up testing.