o
omar_bounawara

Omar B

@omar_bounawara

Python Automation and Data Developer

Tunesië
Engels, Frans, Arabisch
Sommige informatie wordt in het Engels weergegeven.
Over mij
I am a Python developer focused on turning messy data into reliable, structured output. I build web scrapers, scheduled data pipelines and text-to-JSON extractors, and I fix scrapers that keep breaking. I also build tests for LLM apps (RAG, chatbots): test sets, prompt-injection checks and pass/fail reports. My strongest proof is a client data pipeline I built (confidential): daily ingestion, deduplication, rule-based scoring, scheduled runs and crash recovery. One real month run: 318,962 records in about 34 minutes, 0 errors. Based in Tunisia. Arabic, French and English.... Lees meer

Skills

o
omar_bounawara
Omar B
offline • 
Gemiddelde reactietijd: 3 uur

Bekijk mijn diensten

Data scraping
I will fix and harden your broken python scraper or web scraping script
VanafUS$ 60 / projectMeer informatie
Automatiseringen
I will extract structured data from text with an ai llm into validated json
VanafUS$ 80 / projectMeer informatie

Portfolio

Werkervaring

GAAPWise_Accounting

Forward-Deployed Software Engineer (Consulting)

GAAPWise Accounting

Sep 2026 - Sep 2026 • 0 mos

Built a Python pipeline for a client that ingests public French legal announcements daily, removes duplicates, enriches company data, scores leads with rule-based logic and runs on a schedule with crash recovery. One real month run: 318,962 records in about 34 minutes, 0 errors. Delivered, then revised after the client's written review. The client work is confidential.

Institut_Supérieur d'Informatique (ISI)

Research Software Engineer (final-year project)

Institut Supérieur d'Informatique (ISI)

Feb 2026 - Jun 2026 • 4 mos

Designed and built CrawlViz, a focused web crawler, as my solo final-year project. It scores links with local embeddings first and asks an LLM only on selected cases. It runs on asynchronous I/O, with an event-driven backend, a FastAPI control API, live state over WebSockets and a React/D3 view of the crawl. I benchmarked the strategies and wrote the report. Grade: 15.5/20.