Freelancer profile translated to English.
Back to original languageAbout Sameer
I build ML systems that actually work in the real world.
Not proof-of-concepts that die in notebooks. Production systems serving millions of users, 24/7, generating revenue.
Recent Projects :
Computer Vision Systems for Paris 2024 Olympics (3 billion viewers, 99.9% uptime). Improved a retail matching system from 67.8% to 99.4% accuracy, deployed across 21,000 stores, uncovering millions in missed revenue.
What I do :
Computer vision for retail. Document intelligence for healthcare. Object detection for broadcasting. If it involves pixels, tensors, and production infrastructure - I'm in.
Stack :
Python, PyTorch, TensorFlow, Docker, Kubernetes, AWS, FastAPI
The Human Element :
I speak three languages. I can explain neural networks to your CEO and debug CUDA kernels with your engineers. I practice elite calisthenics - if I can hold a planche, I can handle your production pipeline.
Currently :
Available for 3-6 month engagements. Immediate start. Remote or based in Paris.
If you're looking for someone who codes fast, ships even faster, and truly understands what 'production-ready' means - let's talk.
English
Native or bilingual
French
Fluent
Dutch
Basic
Hindi
Native or bilingual
Can work on-site
Paris (up to 50km)
Experience
- MediaPerformancesComputer Vision & Document Intelligence EngineerJune 2025 - Today (1 year)92400 Courbevoie, France
Business Impact :
- Accuracy: 67.8% → 99.4% (+31.6 points)
- Discovered 8,000+ missed opportunities = €millions recovered
- Deployed across 21,000+ stores in production
- 1,000+ req/min, <200ms latency, 99.9% uptime
Technical Stack :PyTorch• Kubernetes (AKS)• OCR• Layout Detection• Auto-labeling• Active LearningAchievements :
✓ Complex data extraction pipeline (tables, various formats)✓ Production CV API serving multiple models simultaneously✓ Reduced downstream errors by 65% via quality gating✓ Saved 70% of annotation costs (1M+ images)Production-ready. Scalable. Immediate ROI. - Olympic Broadcasting ServicesComputer Vision & ML EngineerJune 2024 - September 2024 (3 months)Paris, France
Scale :
- 3 billion viewers served
- **50+ simultaneous video streams, real-time
- 99.9% uptime throughout the competition
- <150ms average latency
Technical Optimizations :
✓ Optimized edge models: 3x acceleration (ONNX/TensorRT)✓ Multimodal search system (CLIP embeddings)✓ Knowledge distillation for edge deployment✓ Real-time scene classification (OpenCV + PyTorch)High availability. Real-time. Zero downtime. - Business Kitz (Software Development)AI/ML Consultant — Multimodal SystemsOctober 2023 - May 2024 (7 months)Brisbane, QLD, Australia
Performance :
- 75% reduction in processing time
- **100,000+ documents/images indexed
- Natural language queries on text+image data
- Scalable multimodal retrieval
Technical Architecture :
✓ Multimodal RAG: LangChain + Vision Models✓ FAISS vector database for efficient retrieval✓ GPT-4 Vision API with optimized prompt engineering✓ Strategic batching & caching for performanceVisual question-answering. Production-ready. Scalable.
Recommendations
Be the first to recommend Sameer
Help this freelancer shine by sharing your experience working together.
These freelancer profiles also match your criteria
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
Education
- Master of ScienceEPITA Paris2025Master of Science
- Bachelor of Technology inReva University2021Bachelor of Technology in
Certifications
- Machine Learning Engineering for Production (MLOps) SpecializationDeepLearning.AI2023
- Production LLM Deployment & Fine-tuningDeepLearning.AI2025