Python Engineer – ML Engineer | Automation Specialist

Ramy Gharib
Driving Operational Efficiency through ML & Automation

Strategic Python Engineer specializing in automation and ML solutions that drive operational efficiency at scale. Expert in the full development lifecycle—from design and testing to deployment and production monitoring. Proven track record of eliminating manual processes and solving complex technical challenges through scalable, maintainable code.

View My Experience View My Top Projects My Resume

$9.6M

Return Cost Reduction

58 HC

Headcount Saved

50%

ASIN Volume Reduction

Areas of Expertise

Languages

Python (Expert), SQL (Advanced), TypeScript, Node.js

Frameworks/Tools

Flask, REST APIs, Selenium, Pandas, NumPy

Specializations

Machine Learning (ML), LLMs, NLP, Computer Vision

Infrastructure

AWS, Linux, Firebase NoSQL, OOP

Key Achievements

Scale & Cost Optimization

Reduced catalogue processing volume from 1,086M to 523M ASINs (50% reduction) by consolidating 8 redundant workflows into a single unified pipeline, while maintaining 100% rule accuracy.

AI/LLM Innovation

Led prompt engineering for an LLM-powered bilingual size detection system, increasing detection accuracy from 75% to 90%+, targeting a reduction of 18,000 customer contacts annually.

Operational Impact

Delivered a projected $9.6M reduction in return costs and $300K in annual net concession savings by resolving Title-Size inconsistencies across 25% of affected Shoes & Apparel ASINs in MENA.

Infrastructure Modernization

Migrated MENA catalogue processing from a legacy framework to a scalable platform, consolidating 8 complex remediation rules into a single Python-based pipeline — improving data actionability by 39%.

Automation at Scale

Built and deployed 11 automation tools across the MENA catalogue quality team, contributing to a total headcount saving of 58 HC and reducing average task time by over 80%.

Catalogue Quality Turnaround

Drove product data quality from 17% to 70% across 300 Product Types — a 270% year-on-year improvement — through automated attribute backfill and enrichment workflows.

Top Projects & Key Initiatives

AI-Powered Bilingual Size Detection (2025)

Role: Prompt Engineering Lead

  • Designed and led prompt engineering for a three-layer LLM-based framework to detect bilingual Title-Size inconsistencies across MENA Softlines (Shoes & Apparel). The system validates English-Arabic consistency, Title-Size attribute alignment, and size system standardization across 4 detection use cases.
  • Enhanced detection accuracy from 75% to over 90%, contributing to an estimated $9.6M reduction in return costs and elimination of approximately 18,000 customer contacts annually.
  • Targeted inconsistency rate reductions across marketplaces: AE (13.3% → 8%), SA (13% → 7.5%), EG (5% → 3%), addressing 25% of affected ASINs in Shoes & Apparel.
  • Projected $300K annual net concession savings alongside the $9.6M return cost target.

Catalog Quality Architecture Redesign (2025)

Role: Lead Engineer

  • Re-engineered 8 legacy catalogue remediation rules into a single unified, sequential Python pipeline with conditional execution logic and cross-marketplace attribute backfilling.
  • Eliminated redundant ASIN ingestion across workflows, resulting in a 50% reduction in annual processing volume (from 1,086M to 523M ASINs) while maintaining 100% rule accuracy.
  • Designed a cross-marketplace integration solution pulling simultaneously from 7 international catalogue data sources (AE, EG, SA, US, UK, IN, AU) to automatically backfill missing attributes and product dimensions.
  • Improved catalogue remediation actionability from 38% to 53% — a 39% gain. Bullet Points Backfill use case improved 230% (26% → 86%).

Customer Service Workflow Automation & ML Integration (2025)

Role: Automation Implementation Lead

  • Customized global Andon workflow frameworks for MENA-specific operational requirements and Standard Operating Procedures (SOPs).
  • Integrated two ML models into the workflow: (1) automated bin check ticket response validation from Fulfillment Centers, and (2) automated vendor response validation — enabling intelligent routing and automated resolution.
  • Partnered with Application Engineering teams, submitting code change requests (CRs) and iteratively refining implementations.
  • Achieved a remediation success rate of 86.9%, exceeding the organizational target of 85%.

Professional Experience

Python Engineer – ML Engineer
Amazon
Jun 2023 - Present
  • Redesigned MENA catalogue quality processing architecture by consolidating 8 independent remediation workflows into a single unified Python pipeline — cutting processing volume from 1,086M to 523M ASINs (50% reduction) while maintaining 100% rule accuracy and reducing infrastructure costs.
  • Designed a cross-marketplace data integration solution pulling simultaneously from 7 international catalogue data sources (AE, EG, SA, US, UK, IN, AU) to automatically backfill missing attributes, descriptions, and product dimensions — improving catalogue actionability from 38% to 53%.
  • Led prompt engineering for an LLM-powered bilingual size detection system across MENA Softlines, designing a three-layer detection framework covering English-Arabic consistency, Title-Size attribute matching, and size system standardization — improving detection accuracy from 75% to 90%+, targeting $9.6M in return cost reduction and $300K in annual net concession savings.
  • Integrated ML models into an automated customer service ticket validation workflow, enabling intelligent routing and automated resolution of bin check and vendor response tickets — achieving a remediation rate of 86.9%, exceeding the 85% business target.
  • Maintained 24/7 production monitoring across all automation pipelines, partnering with QA and DevOps teams to ensure rapid incident response and zero-downtime catalogue data processing across MENA marketplaces (AE, SA, EG).
Automation Engineer
Amazon
Oct 2019 - Jun 2023
  • Built and deployed 11 automation tools across Amazon's MENA catalogue quality team, covering product type classification, data accuracy validation, image quality detection, content enrichment, variation integrity checks, and catalogue coverage allocation — each replacing a previously manual process.
  • Independently researched, trained, and deployed a DistilBERT-based product type classification model fine-tuned on 542 MENA product types, achieving a 91% F1 score on test data — fully automating a catalogue audit process that previously required manual review of ~6,000 ASINs.
  • Automated catalogue audit throughput, reducing processing time from 700 ASINs per 8-hour shift to 700 ASINs in 3.37 minutes, saving an estimated 174 hours per month (~1 FTE), contributing to a total headcount saving of 58 HC across all automation tools deployed.
  • Built a self-service dashboard to detect broken product variations at scale, uncovering that ~40% of ~$60M in high-priority products had broken variations — resolving these defects increased individual ASIN sales by up to 35%.
  • Drove product data quality from 17% to 70% across 300 Product Types — a 270% year-on-year improvement — by building automated attribute backfill and enrichment workflows that eliminated dependency on manual data entry.
  • Developed modular, reusable Python scripts using Selenium and Flask to eliminate repetitive catalogue workflows, reducing average task execution time by over 80% — from approximately 2 hours to under 10 minutes per task.
Senior Operations Coordinator
Amazon
Jan 2017 - Oct 2019
  • Managed and resolved customer escalations with a 99.8% within-SLA resolution rate, directly contributing to a 12% improvement in customer trust scores across MENA operations.
  • Increased product conversion rates by 30% by identifying and fixing catalogue discoverability issues — applying structured content quality improvements and SEO standards at scale across product listings.
  • Enforced SEO and content quality standards across product descriptions, ensuring full compliance with Amazon's website and business publishing requirements across MENA marketplaces.
  • Proactively identified, documented, and escalated operational bottlenecks to senior stakeholders, enabling faster resolution and measurable improvements in team throughput within strict SLA timelines.

Education & Certificates

Bachelor of Commerce (Accounting)

Cairo University

2013 - 2017

Certifications