Semantic search & retrieval
Embedding evaluation, Elasticsearch/OpenSearch, Qdrant, hybrid retrieval, ranking analysis and search relevance improvements.
Typical output: retrieval evaluation, architecture and an implemented search improvement.
Independent engineering consultancy · Berlin · International
AI, NLP & Data Engineering Consulting
Build, evaluate and improve search, machine learning and data systems.
I help technical teams turn difficult data and language problems into testable systems. My work spans production semantic retrieval, Arabic NLP, document intelligence, Python software and ML evaluation.
Ali Abul-Hawa, PhD · 7+ years of industry data-science experience · English & Arabic · Based in Berlin, Germany
What I can deliver
Defined engineering work, not generic AI promises.
Services can be scoped as a technical review, a working prototype, or an implementation milestone.
Arabic & multilingual NLP
Arabic text normalization, classification, dialect-aware processing, language pipelines and multilingual information retrieval.
Typical output: a reproducible NLP pipeline with measurable evaluation.
Document AI & RAG
Document ingestion, metadata-aware chunking, retrieval, grounded Q&A prototypes and offline quality assessment.
Typical output: a tested prototype and evidence about quality, failure cases and next steps.
Python, data & ML systems
Python APIs, SQL/data pipelines, workflow orchestration, automated checks and MLOps components such as experiment tracking and monitoring.
Typical output: maintainable code, validation and a documented handover.
Evidence from real work
Technical case studies you can inspect.
Professional experience, collaborative research and independent prototypes are identified separately. Figures describe the work, not guaranteed client outcomes.
Professional work · Production retrieval
Semantic retrieval for an existing software product
Compared nine embedding variants across 28,400 relation cases before integrating semantic ranking into a multilingual product.
Read the case study →Collaborative professional research · Arabic NLP
Arabic social-language processing at scale
Pipeline work involving more than 700 million raw records, normalization, dialect variation and supervised classification.
Read the case study →Independent prototype · Document intelligence
Local document search and Q&A
Hybrid OpenSearch and Qdrant retrieval, answer provenance, evaluation-driven decisions and observability.
Read the case study →Who you work with
An engineer who can own the technical details.
QortxAI is the independent consulting practice of Ali Abul-Hawa, PhD, an engineer and former senior data scientist with around seven years of industry experience in data science and machine learning.
My background combines electrical engineering, applied ML, Arabic NLP, search relevance, data engineering and hands-on software development. I work remotely with teams internationally, and can discuss scoped engagements with Berlin-based clients.
I do not represent a large agency. Project scope, availability, deliverables and any ongoing support are agreed explicitly.
How engagement works
A practical path from problem to deliverable.
- Describe the problemShare the objective, existing stack and constraints. Do not send credentials or confidential datasets in the first message.
- Agree a bounded scopeDefine deliverables, evaluation criteria, timeline and commercial terms before implementation.
- Build and verifyReview measured results, limitations, documentation and handover material.
Project inquiry
What are you trying to solve?
Send a short description of your use case, technology stack and expected deliverable. I can respond with whether the project fits my expertise and suggest a scoped next step.
Or contact me directly: qortxai@outlook.com · Open email