Based inTehran

Software engineer

I build web products end to end.

Amir Mohammad Jazayeri, software engineer in Tehran

I build web products end to end: the interface, the API behind it, and the PostgreSQL database under both.

I study Computer Engineering at Sharif University of Technology and work as a full-stack engineer at Nisatech. My background is frontend; these days I spend most of my time on backend and AI engineering.

Selected work
installs of Letterboxd Toolkit across Chrome and Firefox
300+installs of Letterboxd Toolkit across Chrome and Firefox
open-source libraries and tools on npm, PyPI and GitHub
4open-source libraries and tools on npm, PyPI and GitHub
live medical-practice websites I maintain
6live medical-practice websites I maintain
course-semesters as a teaching assistant at Sharif
7course-semesters as a teaching assistant at Sharif

Selected work

05 projects

Letterboxd Toolkit store image: the A Life in Film recap, with a marquee title and a ticket showing 1,124 films, 2,031 hours and 584 directors
01 / Browser extension for Chrome and Firefox300+ installs

Letterboxd Toolkit

Native-looking tools for Letterboxd, each switched on separately: a recap of your whole film diary, how the people you follow rated a film, progress rings on lists, and outside scores on film pages.

  • Three tiers: the extension, a stateless API and a cache server.
  • A normalized PostgreSQL replica of TMDB, about 30 tables, queried with typed SQL through Kysely.
  • Private by design: the API only receives film slugs and ratings, and stores nothing.
  • TypeScript
  • Vite
  • Manifest V3
  • Node.js
  • PostgreSQL
  • Kysely
Café Menu on a phone, opening screen: a low-poly coffee cup on a navy and cream checkerboard, with a stream of coffee pouring into itThe menu list: a pinned row of categories, cards for coffee and tea with photos, ingredients and prices, and a bar at the bottom holding two items in the cartThe order code: a QR code to show at the counter, for two items at 280,000 toman
02 / Ordering web app, frontend and APILive demo

Café Menu

A café menu that takes the order as well. Customers browse, fill a cart, and pay online or at the counter with an order code, and staff hand paid orders over from a screen of their own.

  • The hero is a three.js scene written by hand, with no model file: a low-poly cup slides in, a stream pours and fills it, it steams, it is drunk down, and the pour starts over. It loads in its own chunk, runs only while it is on screen, and sits still as a full cup under reduced motion.
  • Built for the thumb: on a phone, cards swipe right to add and left to remove, an item's sheet pulls down to close, and the featured row pans as the page scrolls.
  • The Django API prices every order from the menu, never from the request, and verifies each payment with ZarinPal. A scheduled job settles the payments whose customer never came back.
  • React
  • TypeScript
  • three.js
  • Framer Motion
  • Django
  • PostgreSQL
  • Docker
03 / ESLint plugin and command-line toolOn npm

rtl-doctor

Finds right-to-left bugs in React, Tailwind and CSS code, and fixes the ones that have a single correct answer, without changing how the page looks today.

  • Fixes depend on direction: on a right-to-left page text-left becomes text-end, and inside a dir="ltr" phone input the same class becomes text-start.
  • Checked on a production Persian site: after 18 fixes, 0 of 6,058 elements moved across 9 pages at two widths. Mirroring every class moved 402.
  • Reads the Tailwind version, because the right fix changed between 3 and 4. The rules are tested against real Tailwind 3.4 and 4.3.
  • TypeScript
  • ESLint
  • PostCSS
  • Vitest
The top of a gsc-pulse report for a fictional demo site. The headline reads: Real clicks fell 2.5%. A bot burst added 1,584 clicks that made it look like +28.6%.
04 / Python command-line toolOn PyPI

gsc-pulse

Turns Google Search Console exports into one HTML report: what grew, what dropped and why, and the five things to do next.

  • Four detectors: bot-click bursts, traffic shifts matched to Google's dated updates, CTR gaps, and queries within reach of the first page.
  • Every number is computed from your own exports, and every recommendation shows the formula behind its estimate.
  • The report is one self-contained file that works offline and with JavaScript turned off.
  • Python
  • pandas
  • NumPy
  • Jinja2
Shamsi-Calendar date picker open on Farvardin 1405, with Nowruz holidays marked in red and the 13th selectedpersian-ui form inputs: a national ID field with a validation error, a mobile number, a Sheba number and a price field with a toman and rial switch
05 / Open-source React librariesOn npm

Shamsi-Calendar

A Jalali date picker for React that matches shadcn/ui. It reads your theme's CSS variables, so colours and dark mode work with no setup.

  • No Tailwind needed at runtime, and no configuration: it picks up the theme the app already has.
  • Right-to-left by default, with full keyboard support.
  • React
  • TypeScript
  • jalaali-js
Also

persian-ui

A companion set of React components and utilities for Persian products: inputs for national ID, mobile, Sheba and prices, digit and date helpers, and validators with real checksums.

Experience

  1. 2025 to now

    Software Engineer at Nisatech

    • Build and ship full-stack features across Next.js, Express and PostgreSQL, owning both the interface and the services behind it.
    • Design the REST APIs behind those features, then profile and fix the endpoints that are slowest under load.
    • Work across teams on architecture and code quality, and debug production issues in the frontend and API layers.
  2. 2025

    Frontend Developer, internship at Danesh Solutions

    • Built a real-time chart system that streams live backend data into interactive charts.
    • Helped build an admin panel that sends messages to Telegram, Eita and other platforms, with views for reviewing what was sent.
  3. 2023 to now

    Web and SEO Engineer at Freelance

    • Maintain six live medical-practice websites, owning technical SEO, page performance and content.
    • Ran a Search Console audit across all six and worked through the fix list: crawl errors, indexing gaps, Core Web Vitals.
    • Built client sites from scratch in React and Next.js, including an industrial workshop site and an animated café menu.
  4. 2024 to 2026

    Teaching Assistant at Sharif University of Technology

    • Seven course-semesters, including Web Programming and Software Engineering, twice each.

Education and skills

2022 to 2027 (expected February)

B.Sc. Computer Engineering, software engineering track

Sharif University of Technology

Interests: Software engineering, developer tools, and AI for software engineering.

Languages
TypeScript, JavaScript, Python, SQL
Frontend
React, Next.js, Vue, Tailwind CSS
Backend
Node.js, Express, Django, FastAPI, REST APIs
Data
PostgreSQL, Kysely
Tools
Docker, Git, Claude Code, OpenAI API