Sofien Fejji

Key achievements

• Developed 6+ applied AI applications from concept through implementation, testing and evaluation; verified CI results include 7,500+ passing automated tests across five applications.
• Modernized multi-terabyte data platforms spanning 2,500+ legacy objects and 10+ business units.
• Delivered 100+ dashboards and reports across 20+ business units, serving 1,000+ enterprise users.
• Mentored 6–7 junior consultants; led 10+ knowledge sessions reaching 100+ participants on AI, Data and Cloud.

Experience

1. Senior Consultant, AI and Data Engineering

   Reply Germany, Berlin · Jan 2023 – Present

   • Automated 5 onboarding and knowledge processes across 3 Reply units with LLM and RAG assistants, enabling self-service knowledge access and reducing repetitive support effort.
   • Built governed AI workflows with retrieval, tool calling, structured outputs, guardrails, evaluation, fallback logic and human oversight.
   • Delivered data solutions from ingestion and transformation through modelling, warehousing and BI.

2. Consultant, Data Engineering and Analytics

   Reply Germany, Munich · Jan 2020 – Dec 2022

   • ETL/ELT pipelines and enterprise data warehouses across 3+ client engagements in banking, energy, manufacturing and retail.
   • Cloud migrations and transformation redesign with reconciliation, testing and data-quality controls.
   • Turned stakeholder requirements into data models, KPIs and dashboards.

Education

• AI Engineering Program

  Turing College · Mar – Jun 2026

  • 520 hours, grade 100%. AZAV-certified 13-week program.
  • 4 hands-on projects: LLM apps, RAG, LangChain, agents, APIs, automation, prompt evaluation, capstone.

• B.Sc. and M.Sc. Electrical Engineering and Information Technology

  Technical University of Munich · 2013 – 2019

  • Focus: Data, Analytics and ML.
  • Master's thesis at BMW: deep-learning model for soft real-time driving-style detection, 98% accuracy.

Publication

Modeling Dangerous Driving Events Based on In-Vehicle Data Using Random Forest and Recurrent Neural Network

Co-author · IEEE Intelligent Vehicles Symposium, 2019 · DOI 10.1109/IVS.2019.8814069