Abdulaziz A.
He is creative, very cooperative, and an expert in his field. I will continue to work with him, God willing.
Software engineer in Mersin, Türkiye
Scrapers, bots, and full-stack apps that run in production.
Since 2022 I've shipped 20+ client projects on Mostaql, Fiverr, and Upwork. I keep systems small and data models clean on purpose, so they stay easy to change and hard to break.
/
/_next/static/chunks/main.js
/api/v2/venues?page=1
→ {"total": 24106, "items": [...]}
/api/v2/venues?page=2
Five things clients hire me for most. Everything below runs live in the browser: a bot conversation, code becoming an interface, a scrape from URL list to CSV, a captcha solving itself, a blocked request rerouted through a fresh IP.
Bot development
Bots that answer customers, take files, and call your API behind the conversation. Past builds include a Telegram bot for auto-replies, a Discord bot for a gaming community, and one that pings a buyer when a product is back in stock. Sessions, rate limits, and retries run unattended.
webhooks, long polling, session storage, rate-limit handling
scrapebot
online · replies in seconds
Web and app development
One codebase: React, Next.js, or Astro on the front, FastAPI or Express behind it, Postgres or MongoDB underneath. Recent work includes an AI job matching platform with a custom spaCy model, a public knowledge base for appliance error codes, and a Next.js migration that scored 100/100 on Lighthouse.
React, Next.js, Astro, FastAPI, Express, PostgreSQL, MongoDB
Web scraping
Most sites I'm hired for use logins, infinite scroll, or client-side rendering, so I build for those first: rotating proxies, headless browsers, retry ladders, and parsing that validates every record before it reaches CSV, Excel, or your database. Eight scrapers run in production today, including Amazon, TikTok Shop, Shopify, and Google Maps.
Playwright, Puppeteer, Cheerio, headless Chromium
$
discover sitemap.xml + pagination
1,248 urls across 42 pages
proxy 96 rotating proxies in 12 regions
fetch [ ████████████████████ ] 1,248/ 1,248
js 318 pages return empty html
render headless chromium, 18 tabs
[ ████████████████████ ] 318/ 318
retry 14×429, 3 timeouts, 2 captchas solved
parse JSON-LD + DOM fallback
1,244 records, 4 malformed skipped
clean 13 duplicates dropped, prices normalized
dedupe 1,231 unique by asin
schema asin, title, price, rating, reviews
write deals.csv, 1,231 rows, 412 KB
done 12.4s at 99 requests/s
$
CAPTCHA solving
Image grids, sliders, and token challenges get solved inside the same request flow. OpenCV matching covers image challenges; 2captcha or Capsolver cover the rest. The account pool API I built for a music generation service keeps sessions alive with Playwright and 2captcha, no operator needed.
OpenCV, 2captcha, Capsolver, Playwright
select all vehicles
3 × 3
Proxy infrastructure
A proxy pool sits between the scraper and the site: residential, mobile, and datacenter IPs with health checks, per-domain rotation, sticky sessions for logged-in runs, and failover when an address gets banned. The scraper asks for a page; the pool decides which IP sends it.
residential, mobile, datacenter, sticky sessions, rotation
standing by
connecting via proxy…
403 · access denied
datacenter ip flagged
AirPods Pro 2
$249 4.8
Series 9 · 45mm
$429 4.7
Pixel 9 Pro
$999 4.9
MacBook Air M3
$1,099 4.8
Alpha 7 IV
$2,499 4.9
HomePod mini
$99 4.6
Five projects that show how I work: find the mess, build the smallest system that handles it, then keep it running.
All repositories on GitHubAI job matching
A custom-trained spaCy model pulls skills out of resumes and job posts, and a matching engine scores scraped Indeed listings against them. Candidates get a PDF resume generated through OpenAI, with a retraining endpoint that ships new labels without a redeploy.
resumes + job posts → ranked matches
FastAPI, spaCy, LangChain, React Router v7
Data extraction
An applicant tracking system scraper where each provider is one interface, so adding Greenhouse, Recruitee, or Workday takes a single file. A name-discovery heuristic builds three URL variants per company and finds candidate pages without manual setup.
one interface → every ATS provider
Bun, TypeScript, cheerio, ofetch
API infrastructure
A wrapper around a music generation service with a pool of accounts stored in SQLite, credit-aware routing, and JWT tokens that refresh before they expire. Playwright and 2captcha handle the challenge step, so the API runs unattended behind OpenAPI docs.
account pool → always-on API
Bun, Elysia, Playwright, SQLite
Automation
Eight production scrapers for TikTok Shop, Amazon, Shopify, Google Maps, Mixcloud, Unsplash, and DuckDuckGo. Each one rotates proxies, solves challenges, and dodges bot detection with stealth and OpenCV, then exports clean CSV or Excel files.
eight hostile sites → one spreadsheet
Playwright, Puppeteer, Stealth, Proxy rotation
Knowledge base
A public knowledge base for home appliance error codes. A pipeline crawls DuckDuckGo, extracts structured explanations and fixes with LangChain, validates them with Zod, and stores them in MongoDB for a Next.js site built with NextAuth.
web crawl → structured fixes
Express, MongoDB, LangChain, Next.js
Two internships and four years of freelance work, in the order they happened.
Feb 2026 to May 2026
Eterna Teknoloji, Mersin, Türkiye
Jun 2025 to Jul 2025
Toros Yazılım, Mersin, Türkiye
2022 to present
Mostaql, Fiverr, and Upwork, Remote
The stack I reach for. I pick tools that stay boring under load.
Reviews from Mostaql and Fiverr, all five stars. Repeat clients show how many projects we did together. Every card links to the original review, and hovering one shows the full text.
Abdulaziz A.
He is creative, very cooperative, and an expert in his field. I will continue to work with him, God willing.
Alia E.
I was very happy to deal with you. Very cooperative and fast. He deserves more than five stars; I recommend him without hesitation.
Abu Hazem A.
10 projectsThis is the second time I have worked with him. He is an honest, truthful, and very cooperative man. God willing, our cooperation will continue.
Abu Hazem A.
10 projectsA very cooperative man and honest in his dealings. This is the third time I have worked with him, and I will not hesitate to deal with him in the future, God willing.
Khaled A.
Dealing with this person was extremely professional. He is very cooperative and open to all feedback. Thank you from the heart.
wttheelen
Abdulghani was very helpful. I had a project I wanted done and my initial brief was a bit too vague to start but he he had a good intuition of what I was trying to get at and asked the right questions to get started. After that the project went well, responses were prompt even when I said I wasn't in a hurry, I am quite happy with the results of the work and will continue to ask with his help for future projects.
Faleh M.
Fast execution and proficiency. I recommend dealing with him.
haituni2
3 projectsAs always very good
Abu Hazem A.
10 projectsMy approved/go-to programmer on Mostaql.
Abu Hazem A.
10 projectsMastered work performed to the fullest extent.
Abu Hazem A.
10 projectsA very polite and helpful man. He provided the required service perfectly without any issues. God willing, we will have future cooperation without hesitation. Five stars for all criteria.
haituni2
3 projectsgood and perfect
nodgefan86
excellent fast delivery :)
haituni2
3 projectsvery good and fast service
Abu Hazem A.
10 projectsI am still continuing to request projects from this man because of his ethics, advice, and professional dealing.
Abu Hazem A.
10 projectsThe freelancer I trust the most for the quality of his work.
Abu Hazem A.
10 projectsThe freelancer I trust the most; a sincere and honest man in his work, God willing.
Abu Hazem A.
10 projectsA sincere and very cooperative man. I will not hesitate to deal with him.
Abu Hazem A.
10 projectsGod bless you; honest and sincere.
Send me a short description of what you need: what the data is, where it lives, and what you want out of it. I reply within a day.
Based in Mersin, Türkiye. Open to freelance work and full-time roles.