À propos de Silas
I help companies turn AI, LLM, and Computer Vision into production-ready solutions that drive measurable business impact. AI Engineer with 5 years of experience delivering end-to-end AI systems across healthcare, retail, finance, and smart-city domains, specializing in designing, optimizing, and deploying scalable AI/ML/LLM/Generative AI solutions from research to cloud integration. I bridge cutting-edge AI with practical outcomes faster processes, lower costs, better decisions, and new product capabilities delivering systems used across multiple countries, including medical imaging AI, Document AI, recommendation engines, and real-time vision systems.
Specialization:
- Build and deploy custom AI / ML / LLM / Generative AI solutions
- Computer Vision: detection, segmentation, OCR, tracking, medical imaging
- NLP & LLM apps: RAG, AI agents, document understanding, chatbots
- Multimodal AI and recommendation systems
- Model optimization and edge/cloud deployment (Docker, APIs, Azure/AWS/GCP)
- Productionization, MLOps, and performance tuning
Typical projects & deliverables:
- Document AI / OCR pipelines for automated data extraction
- LLM or RAG applications for enterprise knowledge and automation
- Computer vision systems for images, video, or medical data
- Generative AI content or recommendation engines
- End-to-end AI APIs and microservices ready for integration
- Optimized models for real-time or cost-efficient inference
Anglais
Bilingue ou natif
Français
Notions
Accepte de travailler sur site
Paris (jusqu’à 50 km), Lyon (jusqu’à 10 km), Marseille (jusqu’à 10 km), Bordeaux (jusqu’à 10 km), Toulouse (jusqu’à 10 km)
Expériences
- Roc4TechAI EngineerAGENCE & SSIIdécembre 2025 - mars 2026 (3 mois)
- Developed and deployed a retail queue management system that uses YOLOv9 and Norfair to track customer flow from cameras and XGBoost to forecast lane demand. System was deployed on GCP using Cloud Run for scalable inference APIs, Vertex AI for model management, and BigQuery for analytics and storage, enabling real-time insights that help managers proactively open lanes and reduce wait times.
- Tribun Health,AI EngineerBIOTECHNOLOGIESfévrier 2025 - octobre 2025 (8 mois)Paris, France
- Designed and deployed a multi-species mitosis detection system using PyTorch, Detectron2, Docker, and Flask API on Whole Slide Images from multiple scanners. Combined Detectron2-based detection with DenseNet and EfficientNet classifiers for cell classification, outperforming 80% of public pathology AI benchmarks and enabling 3× faster slide review for pathologists; ViT models were evaluated but underperformed.
- Fine-tuned a Triple-Head HoverFast-based cancer cell segmentation and quantification system in PyTorch using teacher–student distillation and Macenko stain normalization to adapt from H&E to IHC images, achieving 91% correlation with expert-annotated ROIs. Deployed the solution using Docker with a Flask-based inference API to support clinical workflows.
- Anheuser-Busch InBevAI EngineerVINS & SPIRITUEUXfévrier 2024 - janvier 2025 (11 mois)Lagos, Nigeria
- Designed and deployed a speech-to-text and LLM-powered audio system using Whisper, OpenAI GPT, categorizing and routing 10,000+ monthly customer queries, reducing manual triage by 70% and improving response times by 35%.
- Designed and developed a Document AI system using Detectron2 for ROI detection and fine-tuned PaddleOCR for text extraction, automatically extracting and categorizing records from 15+ global retail subsidiaries, reducing processing time by 70% and improving data accuracy and compliance.
- Optimized the Document AI pipeline using OpenVINO, reducing model size by 96% through quantization and improving inference speed by 35%, cutting cloud costs and virtually eliminating data processing backlogs.
- Developed an LLM-based Generative AI recommendation engine using Open AI API and deployed across Africa and South America, generating localized ad content in multiple languages and dialects, driving a 45% increase in customer engagement and 27% uplift in conversion rates.
- Collaborated with product and marketing teams to align AI insights with business objectives, integrating the system into the company's global retail ecosystem.
Recommandations
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
Formations
- MSc. Health ArtificialUniversite de Bourgogne Europe2025MSc. Health Artificial
- BSc. Industrial PhysicsFederal University of Agriculture2021BSc. Industrial Physics
Certifications
- Edge AI CertificationIntel2022
- Data analytics CertificationGoogle2022