Hire the Best Machine Learning Engineers
in the Netherlands

Innovative Machine Learning Engineer and Science Educator with a unique dual expertise in developing advanced AI/ML-solutions and translating complex scientific/technological concepts for diverse audiences. With over 20+ years professional experience, of which 8+ years in technology and engineering, I specialize in building custom AI, Machine Learning, and NLP solutions that solve real-world challenges (such as high-accuracy computer vision models (YOLO), optimizing predictive analytics (XGBoost/PyTorch), or leveraging the latest LLMs for automating workflows). In addition to that, I am also an expert in crafting highly accessible course materials about a plethora of topics from science and technology. My technical foundation is rooted in rigorous mathematics and statistics, but my work isn't just theoretical; I have developed my skills in the world’s most competitive environments, including Kaggle, where I’ve mastered the full Python data science stack. Key Competences: - Deep Learning & CV: YOLO, PyTorch, Keras, TensorFlow, XGBoost. - Advanced NLP: LLM prompt engineering (Google/Innodata), BERT, and HuggingFace. - Data Science: Regression, tabular data analysis, and complex time-series prediction. - Course and curriculum development: Science, physics, mathematics, engineering, computer science, cryptography, AI/ML. - Training delivery: Data governance and agentic AI for executives. As a long-term digital nomad, I bring a proven track record of managing complex projects independently and delivering on schedule, regardless of geography.

Data Scientist with expertise in machine learning models, AI research, scalable data pipeline, and data-driven solutions; developed a demand forecasting model achieving 91% accuracy. Proficient in diverse data science tools and techniques, including NLP, LLM, PySpark, and Azure ML Studio, CI/CD, and deploying ML solutions with end-to-end ownership research environments. Collaborated with clients to implement data-driven solutions that improved operational efficiency.

I’m an AI developer and consultant with a strong background in data-driven problem solving, model development, and technical education. I help startups, researchers, and small businesses turn ideas into working AI and data solutions — from early prototypes to deployable systems. My experience spans machine learning, reinforcement learning, and LLM customization, using tools like Python, PyTorch, TensorFlow, and AWS. I also hold industry certifications and have served as a Teaching Assistant, giving me a strong foundation in communicating complex topics clearly and managing collaborative technical projects. Whether you need to explore a concept, analyze data, or integrate intelligent automation, I provide end-to-end support — from scoping and design to implementation, documentation, and structured progress updates. I value transparent communication and async-friendly workflows that keep projects moving smoothly.

Expert-vetted top 1% AI/ML Researcher at UpWork. I am an experienced researcher/ engineer in the domain of applied machine learning. I have 10+ years of experience in the field of AI/ML, have a PhD in deep learning, worked at Apple for 4 years, and a cutting edge AI-hardware startup for 2 years. I have worked across different problem domains: computer vision, information retrieval, GPT/ NLP, custom-GPT models, object detection, 3D pose estimation, etc. After working on these different domains over the past 10 years, I have come to appreciate designing ML pipelines end-to-end for a particular problem. I believe designing well thought ML pipelines end-to-end wins over standard out of the box ML models. More details about my past: I was employed as a principal research scientist at Cerebras, California, USA. My work involved doing hardware aware machine learning research on accelerating training and inference of neural networks. Prior to working at Cerebras, I was employed as a research scientist within special project group (Self-driving/ Autonomous vehicles) at Apple. I conducted research at the intersection of robotics, computer vision and machine learning. Prior to joining Apple, I was a graduate student at INRIA Grenoble, France. During my PhD, I worked on data and method driven machine learning models for architecture learning of neural networks, image segmentation, verification and retrieval and dimensionality reduction methods. The fundamental contribution of my PhD was work on "Convolutional Neural Fabrics", a method for learning the architectures of neural networks in an end-to-end manner.

$50k ROI | 6+ Years Experience | Delivered for USPS, Airbus, Saudi Post, and Defense Clients I help enterprises bridge the gap between "cool AI demos" and production-grade systems that scale. Whether it's deploying 200+ real-time cameras in industrial environments, fine-tuning LLMs for 23x faster inference, or architecting Agentic workflows with LangGraph, I deliver high-performance solutions that drive measurable business ROI. 🚀 Recent High-Impact Wins Industrial IoT & CV (Sigrow): Architected a real-time plant health monitoring system deploying 200+ multispectral/thermal cameras, ingesting 2.3M images/month with YOLO-based segmentation and visual-thermal data fusion. LLM Optimization (Rapidev): Fine-tuned LLaMA 3.2-8B (LoRA/QLoRA) for enterprise NLP, achieving 23x faster inference and automating internal workflows via LangGraph/ReAct agents (50% reduction in manual effort). High-Scale AI Pipelines: Built an inference engine handling 100K+ daily requests (OCR, Speech, Translation), resulting in $50,000 annual cost savings. Edge Performance: Optimized IoT camera management by migrating from Python/Raspberry Pi to GoLang/MIPS, reducing bandwidth and operational costs while increasing API response speeds by 5x. Defense & Satellite: Built satellite object detection (YOLT/Solaris) with 95% accuracy on small targets (<10px), reducing manual inspection time by 65%. 🧠 Core Expertise Agentic AI & GenAI: LangChain, LangGraph, OpenAI, DeepSeek, LLaMA 3.x. Expertise in ReAct agents and multi-agent orchestration. Computer Vision (Edge & Cloud): YOLO (v5-v11), DeepStream, TensorRT, Triton Inference Server, GStreamer, 3D CNNs. Backend & ML Infra: Python, GoLang (specialized in high-performance IoT/MIPS), FastAPI, Docker, MQTT, RTSP, AWS (SageMaker, Bedrock, EC2). Data Fusion: Integrating multispectral, thermal, and sensor data for actionable industrial insights. 🏆 Why Work With Me? ✅ Gold Medalist & Researcher: MS in Computational Science (NUST) and published author in Deep Learning. ✅ Open Source Contributor: Contributed to TensorRTX and YOLOv7-Pose; PyCon Speaker. ✅ Full-Stack Ownership: I don't just train models; I build the GoLang backends, the Dockerized microservices, and the CI/CD pipelines to keep them running. ✅ Business-First Mindset: I focus on KPIs—whether it's a 60% revenue increase for a Japanese casino or a $50k reduction in API costs. 💬 Ready to scale your AI product? Let’s discuss your architecture and how we can build a scalable, production-ready solution together.

KVK: 95634347 Experienced Machine Learning Engineer with a unique proficiency in diagnosing and resolving user common issues, backed by hands-on experience as an official support engineer. Deep understanding of internal systems ensures efficient problem-solving and tailored solutions to meet user needs. Certificate - Google Cloud Professional Data Engineer - Google Cloud Professional Machine Learning Engineer - AWS Certified Solutions Architect – Associate - Neo4j Certified Professional As an official Google Cloud Platform support Subject Matter Expert(SME), I am well versed in building and maintaining machine learning systems using cloud platforms such as Google Cloud/AWS. - Expertise in diagnosing user-centric issues derived from an official support engineer role. - Comprehensive understanding of internal systems to provide quick and efficient solutions. - Ability to find the customer's pain spots Previously worked as Computer Vision Engineer at deep tech startup in Japan where I clean/prepare/model/validate/test data to deploy deepsort(Object Tracking) into the edge device(NVIDIA Jetson) by converting the configuration file used for the core algorithm with the open-source memory-less framework. ML Tech stack: Python/R/Tensorflow/Pytorch/OpenCV/Spark/Hadoop/Docker/Kubernetes/AWS/GCP/Apache Airflow/Apache Beam/Kubeflow I am confident that my skills and overall knowledge of Cloud support give you not the fancy overengineering but Industry standard solution as is. Link: babushkai.com *PS: you might be able to find me somewhere with kuwabara@google.com