Hire the Best Machine Learning Engineers
in Spain

Building a custom computer vision application? Need an AI model that actually performs in the real world, rather than just looking good on a testing dataset? Hi, I’m Gebrecherkos. I help startups, researchers, and hardware teams design, train, and deploy high-performance Deep Learning and Computer Vision pipelines. Because my background is heavily rooted in autonomous robotics, I don't just build models—I build robust, lightweight systems that are designed to survive messy, real-time edge deployments. Whether you need to detect anomalies in medical scans, build a custom object detection pipeline, or optimize a neural network for real-time inference, I can bring your AI project to life. 🧠 CORE MACHINE LEARNING & CV CAPABILITIES: 🔹 Computer Vision & Image Processing: End-to-end pipeline development for Object Detection, Semantic Segmentation, and Image Classification. Building and fine-tuning custom CNN architectures (e.g., EfficientNet, ResNet) and U-Net models for binary/multiclass segmentation. Expertise in extensive data augmentation strategies for domain adaptation (blur, jitter, synthetic generation) to ensure models generalize to the real world. 🔹 Advanced Deep Learning Architectures: Designing complex Multimodal Data Fusion systems. Developing and implementing Graph Neural Networks (GNNs) for highly specialized classification tasks (e.g., medical skin lesion classification). Deep expertise in PyTorch and optimizing custom loss functions (Cross-Entropy, Lovász-Softmax) for edge-case accuracy. 🔹 Real-World AI Deployment (The "Robotics" Edge): Unlike pure data scientists, I know how to make AI interact with the physical world. Experience deploying ML models onto edge devices and integrating them with autonomous navigation systems (ROS1/ROS2, SLAM). Proven track record with complex datasets, including underwater Unexploded Ordnance (UXO) detection and highly imbalanced medical imaging. 🛠️ TECH STACK: PyTorch, Scikit-learn, Pandas, OpenCV, Matplotlib, ROS1/ROS2, Linux, Git. Let's build something intelligent. Send me a message with a brief description of your Computer Vision or Machine Learning roadblock, and I will provide a high-level technical approach on how we can solve it—no strings attached.

Experience: Developed and deployed end-to-end machine learning pipelines using Azure Machine Learning (AML) and MLflow for model tracking, versioning, and reproducibility. Trained models at scale using Azure ML compute clusters, leveraging AutoML, Python SDK, custom scripts and pipelines for experimentation and efficiency. Designed Retrieval-Augmented Generation (RAG) on-premises / Azure / AWS Developed infrastructure-as-code (IaC) templates using Terraform and GitOps with FluxCD Provided technical expertise in containerization technologies (e.g., Docker, Kubernetes) for deploying microservices-based applications on the cloud platform and on-premises data center Designed and implemented CI/CD pipelines using Azure DevOps and GitLab, enabling streamlined deployments and improved release cycles. Managed and configured Azure Entra ID (formerly Azure Active Directory) for secure identity and access management across cloud resources. Developed monitoring stack using Prometheus, Grafana, Thanos, Telegraf for proactive system health checks

🏆 100% JSS | Top Rated ⏱️ Available across US, Canada, Europe, UK, Australia, and preferred time zones 🥇 Experience with teams at Turing, Meta, Google, YouTube (Agentic) 💼 8+ years | 22+ delivered projects | 506+ billed hours Senior AI/ML Engineer LLMs, RAG, Multi-Agent Systems, Python Backends Expert| Deep Learning | Computer Vision I’m Syed Mohsin Ali Shah, Ph.D. (Computer Science), University of Deusto, Spain. I design and deploy production AI that combines Large Language Models, retrieval-augmented generation (RAG), and agentic orchestration with scalable Python services. Focus areas: measurable accuracy, latency reduction, reliability, and maintainability. Core Strengths .) AI & LLM Engineering: GPT-4/5 class, Claude, LLaMA, Mistral; HF Transformers; custom inference servers .) Agentic AI: LangGraph, LangChain, MCP; planner–executor tool use, guardrails, monitoring .) Agentic AI: LangGraph, LangChain, MCP; planner–executor tool use, safeguards, and monitoring .) RAG: Pinecone, Weaviate, Chroma, LlamaIndex; hybrid search, reranking, retrieval evaluation .) NLP: Summarization, NER, structured extraction, document QA, routing Backend & Platforms: .) APIs: FastAPI, Django/DRF, Flask; JWT/OAuth, rate limiting, observability .)Scale & Ops: Docker, CI/CD (GitHub Actions), Redis, Celery, async I/O, streaming .) MLOps: MLflow, experiment tracking, canary releases, automated eval harnesses Machine Learning, Computer Vision & Generative AI: .) Fine-Tuning: LoRA/QLoRA, SFT, preference optimization; prompt/system design .) Vision: YOLO pipelines, OCR (EasyOCR, Tesseract, PaddleOCR), document AI .) Image Gen: SDXL/Flux, DreamBooth, control adapters, safety layers Cloud Based Deployment: .) Cloud: AWS, GCP, Azure, RunPod; GPU hosting, autoscaling, cost control .) Pipelines: ETL/ELT, Spark, batch/stream retrieval layers, vector ETL Quant & Time-Series (Stocks/Crypto): .) Forecasting Models: LSTM/GRU, TCN, Temporal Fusion Transformer, N-BEATS, XGBoost/LightGBM hybrids .) Signal Engineering: multi-horizon forecasts, regime detection, volatility features, market microstructure signals .) Back testing: walk-forward validation, cross-asset/rolling windows, slippage/fees modeling, position sizing .)Deployment: real-time inference APIs, feature stores, event-driven triggers, risk constraints and monitoring .) Use Cases: stock and crypto price prediction, alpha signal research, anomaly detection, and alerting (Research and engineering services only; not financial advice.) What I Deliver: .) Agentic AI platforms (planner/orchestrator + tools: web search, CRMs, booking, calendars) .) RAG chatbots over PDFs, Notion, websites, and email—evaluated for grounding and faithfulness .) High-performance Python backends with FastAPI + Redis for low-latency inference .) LLM cost/latency optimization with caching, prompt shaping, and model routing .) Quant research stacks for equities/crypto: forecasting, signal evaluation, and live endpoints Tooling & Stack: .) Python, FastAPI, Django/DRF, LangGraph, LangChain, MCP, LlamaIndex, HF Transformers, Pinecone, Weaviate, Chroma, MLflow, Redis, Celery, Docker, GitHub Actions, SDXL/Flux, YOLO, OCR toolkits. .) Time-series: PyTorch/Lightning, darts, sktime, prophet (baselines), vector DBs for event retrieval. 💬 Let’s Work Together I build intelligent, reliable, and scalable AI systems from custom GPT apps and multi-agent workflows to high-performance Python APIs with clear communication, fast delivery, and long-term maintainability.

(Available: October 2025) Stop wasting time on repetitive tasks. I build AI automation that works 24/7 so you don’t have to. I build intelligent automation systems that don’t just save time—they revolutionize business operations. With 7+ years of experience, I design and implement AI-driven solutions that turn slow, manual processes into ultra-efficient, scalable workflows. 🌟 Services: - End-to-end process automation - AI/ML integration - Custom dashboard development - Business process optimization - Data pipeline creation - System integration 📈 Key Achievements: 🟢 AI-Powered Document Processing System - Reduced processing time from 20 min to 0.49 sec with 95.7% accuracy, automating 300+ invoices/month. 🟢 Automated Content Generation - Cut production time from 4 hours to 15 min with multi-LLM integration and seamless CMS automation. 🟢 Real-Time Performance Analytics - Automated reporting saving 2-3 hours daily with custom KPI dashboards & trend forecasting. 🌟 Tools & Tech: -Languages: Python, SQL -AI/ML: OpenAI API, Google Document AI, Computer Vision -Cloud: AWS, Google Cloud -Automation: Make.com (Integromat) -Data & Dashboards: Looker Studio, Google Sheets API -DevOps: Docker, Git, Jenkins Feel free to reach out and we can start right away!

I have a degree in Telecommunications Networks Engineering and also possess a minor in Data Science, both gained in university Pompeu Fabra. I have 8 years of experience with Python and about 2 years of experience with LLM. I have experience with RAG and Langchain, development of AI agents and FastAPI. I have extensive experience building end-to-end RAG pipelines. In my recent work at Revolut as an IAM InfoSec, I developed a data pipeline using the Jira API to extract access-request tickets, processing them via LLM-driven analysis (LangChain) for automated reporting. I am comfortable with the entire lifecycle: corpus preparation, embedding generation, and context assembly. I am a strong advocate for "clean code" (typing, testing, and packaging). My experience as a DevOps Intern at Arbolus and a SysOps Engineer at Vendo Services has given me a deep proficiency in Docker, CI/CD (GitHub Actions), and AWS (ECS, RDS, S3), ensuring that the AI applications I build are scalable, observable, and cost-effective. I have experience with auditing the ISO27001 standards which I gained during my stay at AQU Catalunya as well as IT Support at Arbolus. Also, as a SysOps at Vendo, I ensured our company not only to be compliant with ISO27001 standards but also with PCI DSS and GDPR standards. Also, my experience developing Ansible automation and complex configuration scripts aligns perfectly with your requirement for orchestrating services and managing back-office workflows via tools like n8n or Airflow. I did 2 personal projects: In the first one, I have successfully developed autonomous AI agents using LangChain and integrated them into conversational platforms like WhatsApp and Telegram via APIs. Also, I engineered a script for stock market prediction, trained it on Apple stocks and achieved an accuracy rate of 98.5%. I have a C1 level in English and Spanish. I also speak Russian, Catalan, Azeri and a bit of Ukrainian. You can find more information about my previous experience on LinkedIn and also feel free to check my Github (SH-SW). Fun fact, I play chess since I was 4. Being the national champion back in 2014-2015, my highest achievement was getting into top 10 of Europe in youth, with FIDE ELO 1861. Also, my experience developing Ansible automation and complex configuration scripts aligns perfectly with your requirement for orchestrating services and managing back-office workflows via tools like n8n or Airflow. I did 2 personal projects: In the first one, I have successfully developed autonomous AI agents using LangChain and integrated them into conversational platforms like WhatsApp and Telegram via APIs. Also, I engineered a script for stock market prediction, trained it on Apple stocks and achieved an accuracy rate of 98.5%. I have a C1 level in English and Spanish. I also speak Russian, Catalan, Azeri and a bit of Ukrainian. You can find more information about my previous experience on LinkedIn and also feel free to check my Github (SH-SW). Fun fact, I play chess since I was 4. Being the national champion back in 2014-2015, my highest achievement was getting into top 10 of Europe in youth, with FIDE ELO 1861.

Building LLM-powered products or generative AI systems? I specialize in LLMs, RAG, and image generation (GANs, Diffusion, NeRFs). What I deliver: - Built AI products for YCombinator startups and Fortune 500 companies - Delivered production ML systems in San Francisco, London, and Madrid - 10+ years in AI/ML, from cutting-edge research to production Specialties: LLMs & Language: RAG systems, fine-tuning, prompt engineering, chatbots Generative Vision: GANs, Stable Diffusion, image-to-image translation, NeRFs Production: End-to-end ML pipelines, from data to deployed API Stack: PyTorch, TensorFlow, Hugging Face, Langchain, Diffusers, Docker, AWS/GCP Credentials: Ph.D. in AI | MBA | Expert-Vetted (Top 1% on Upwork) Let's discuss your project - message me to start.