À propos de Outman
- Rebuilt legacy document systems at DXC Technology into intelligent, cloud-native platforms.
- Designed recruitment intelligence at Africawork, using OCR + LLM pipelines to parse CVs and match candidates faster than humans ever could.
- Delivered AI image generation for ads, rule-based and ML recommender systems, and retrieval-augmented search that actually works.
- AI engineer who speaks business, not just code
- Prompt engineer who knows how to make LLMs useful
- Computer vision solutions that see what humans miss
- MLOps pipelines that don’t break in production
Anglais
Bilingue ou natif
Français
Bilingue ou natif
Expériences
- DXC TechnologyAI Engineer + MLOps EngineerAGENCE & SSIIfévrier 2025 - Aujourd'hui (1 an et 4 mois)Most document systems fail for one reason:they don’t scale when reality gets messy.Documents grow.Pages multiply.Layouts change.Forms overlap.Data hides in places rules can’t reach.My mission at DXC was to redesign the platform so it could think structurally, scale effortlessly, and perform reliably, no matter how complex the documents became.I worked on transforming a fragile, rule-heavy system into a modern AI-driven architecture built for speed, accuracy, and long-term evolution.The goal wasn’t to extract text.The goal was to understand documents at scale.
What I Built
- A scalable AI architecture capable of processing large multi-page documents
- A configuration-driven extraction system with zero hard-coded logic
- Dynamic prompt generation driven entirely by structured schemas
- Strong separation logic to prevent data leakage between similar document types
- Hybrid intelligence combining OCR precision with LLM reasoning
- Production-grade pipelines designed for reliability, traceability, and growth
Technology Stack
Cloud & Infrastructure- AWS (Lambda, Bedrock, S3, Step Functions, DynamoDB, CloudWatch)
- Docker
- Event-driven architectures
AI & NLP- Large Language Models
- Advanced prompt engineering
- Schema-based extraction
- Multi-document reasoning
- OCR + LLM hybrid pipelines
MLOps- Prompt versioning and lifecycle management
- Environment isolation (dev / staging / prod)
- Automated evaluation pipelines
- Confidence scoring frameworks
- Observability and performance monitoring
Backend & Data- Python
- JSON schema design
- Asynchronous processing
- High-volume batch orchestration
- AfricaworkAI Engineerjuillet 2020 - janvier 2025 (4 ans et 6 mois)At a recruitment-focused company, I worked on designing the backend intelligence behind systems that connect people to opportunity.I built data-driven services that transform unstructured resumes into structured, usable information, and then use that data to intelligently match candidates with relevant job offers.The work combined classical machine learning with modern AI.Not just automation, it was a full decision-making pipeline.
Stack & Focus
- Python backend systems
- Machine learning–based matching algorithms
- LLM + OCR pipelines for CV parsing
- Hugging Face translation models for French to English resume translation
- Structured data normalization, scoring, and enrichment
The result: faster hiring, smarter recommendations, and systems that scale as resumes grow.
Recommandations
Soyez le premier à recommander Outman
Contribuez à la réussite de ce freelance en partageant votre expérience de collaboration avec lui.
Ces profils de freelance correspondent également à vos critères
Agatha Frydrych
Backend Java Software Engineer
4.7
(3)
2
Baptiste Duhen
Fullstack developer
4.6
(4)
5
Amed Hamou
Senior Lead Developer
4
(2)
7
Audrey Champion
Web developer
4.3
(3)
4